Awesome Virtual Cell

English · 简体中文

Papers, datasets, benchmarks, and community resources for AI-powered virtual cell research.

Virtual Cells  ·  Foundation Models  ·  Perturbation Modeling  ·  Multimodal & Spatial  ·  Intervention Design

_{Start Here · Latest Updates · Papers · Datasets · Challenges}

🚀 Start Here

New to virtual cell research? Pick a track instead of reading the full list from top to bottom.

Start with the bold entry in your track, then compare the other examples. These are selective reading suggestions, not a leaderboard; see the full catalog for publication status.

Goal	Recommended entry points
🧬 Understand the field	Cell perspective · Grow AI Virtual Cells · Nature perspective
🧫 Predict perturbations	GEARS · STATE · PIE · MAP · CellFlow
🧠 Cellular foundation models	Geneformer · scGPT · scFoundation · UCE · TranscriptFormer
🌐 World models & cell-state transitions	A world model of the virtual cell · CellOS · Chreode
🖼️ Multimodal & spatial	Nicheformer · VirTues · DePass · UniPert-G2CP
🎯 Intervention design	PDGrapher · DrugReflector · LabCompass · CellNavi · PAIRING · VCDesign
📏 Evaluation & measurement	Systema · SCMBench · PertResolve · Signal, Bounds & Baselines · Principled Evaluation

📰 Latest Updates

Deadlines for the live competitions first, then a dated log of what changed in this list. Competition entries move into the log once they close.

Live challenges (dates as published by the organizers, checked on 27 October 2026)

Challenge	Phase now	Next milestone	Final call
Virtual Embryo Challenge (NeurIPS 2026)	P3 test phase from 27 Oct 2026, 09:00 PDT (16:00 UTC), a week later than first scheduled: the default leaderboard switches to Test, P2 standings freeze, and eligible teams can request the P2 validation answers	Final submissions due 2 Dec 2026; official evaluation on the hidden test sets starts 4 Dec 2026	Winners announced 11 Dec 2026 at NeurIPS
Virtual Cell Challenge 2026 (Arc Institute)	Final phase since 22 Oct 2026: the final test set (controls-final, contexts D–F, a new 300-perturbation CRISPRi panel) is out, the validation leaderboard no longer takes entries, and final-entry scores stay hidden	Final submissions due 5 Nov 2026, 23:59 UTC (vcc CLI 0.3.0 or newer)	Winners and the final leaderboard announced late Nov 2026

Updates (one line each, newest first; full entries in the update log)

Date	What changed
2026-10-28	Stricter inclusion rules (venue allowlist, preprint bar, technical reports); 12 NeurIPS 2026 papers added; weekly curation workflow
2026-10-27	Literature sweep: 155 papers added (396 to 551) and 9 preprints upgraded to their published venues
2026-10-26	Open community baseline with code for the Virtual Cell Challenge 2026 (community PR)
2026-10-13	Literature audit: 67 re-verified papers added (329 to 396), DOI fixes and venue upgrades
2026-10-10	LabCompass added, with official code, workflows and data

Earlier updates (18)

Date	What changed
2026-10-02	42 source-checked papers added (285 to 327); new Start Here reading tracks
2026-10-01	Two perturbation atlases added to Datasets; Virtual Cell Challenge 2026 dates updated
2026-09-26	New Evaluation & Measurement branch (PertResolve, Signal–Bounds–Baselines, VCBench and more)
2026-09-26	Intervention Design expanded (VCDesign, PDGrapher, PAIRING, PerturbNet, ARC, NUDGE)
2026-09-26	Full Research Papers list restored; all 275 records reviewed for topics and code links
2026-09-25	September sweep (Speciesformer, PHAROS, LucaCell and more); STATE, Tahoe-100M and scBaseCount upgraded to Cell 2026
2026-08-21	Virtual Embryo Challenge added; both competitions grouped; this log started
2026-08-17	World Model and JEPA papers added; preprint links and venue labels refreshed
2026-08-10	Virtual Cell Challenge section covering both Arc editions
2026-08-09	Single-cell and protein papers (scTranslator, CAPTAIN, scPROTEIN)
2026-07-16	TCGA and HEST Xenium virtual spatial transcriptomics datasets (community PR)
2026-07-14	DeepSpot2Cell (NeurIPS 2025) (community PR)
2026-07-06	27 new 2026 papers from a May to July sweep
2026-07-04	DeSCOPE (community PR)
2026-06-22	2026 papers and missing code links
2026-05-28	Datasets section expanded; recent virtual cell papers
2026-05-13	Automated literature update workflow (manual dispatch)
2026-03-18	README restructured; contributing guide added

🔬 Research Papers

Browse the full research collection below; each paper may carry multiple independently assigned tags.

552 papers · 2026 (402) · 2025 (123) · 2024 (27)

Papers can have multiple topics. Search and combine topics · Tag definitions

🗓️ 2026 — 402 papers

• [Predicted-NN-distance] [Perturbation] [Foundation Model] [Evaluation & Measurement] A low-rank decoder bottleneck bounds reliability in foundation-model perturbation prediction (NeurIPS 2026) [paper]

• [NOVA] [Morphology] [Perturbation] [Representation Learning] Biological Graph Priors Enable Representation Learning for Cellular Microscopy (NeurIPS 2026) [paper]

• [PT-RAG] [Perturbation] Differentiable Retrieval-Augmented Generation for Predicting Cellular Responses to Gene Perturbation (NeurIPS 2026) [paper] [preprint] [code]

• [DynaCell] [Morphology] [Benchmark] [Dataset] [Perturbation] [Evaluation & Measurement] DynaCell: an Evaluation Framework for Dynamic 3D Virtual Staining of Live Cells (NeurIPS 2026) [paper]

• [SPEAR] [Perturbation] Spectral Adaptive Repositioning for Flow-Based Single-Cell Perturbation Modeling (NeurIPS 2026) [paper]

• [scOntoFM] [Foundation Model] [Representation Learning] Learning Biological Hierarchies in Single-Cell Foundation Models (NeurIPS 2026) [paper]

• [Regulator Tokens] [Spatial] [Morphology] [Multimodal] [Gene Regulation] Pathway-Aligned Regulator Tokens for Interpretable Spatial Gene Expression Prediction from Histology (NeurIPS 2026) [paper]

• [scIFLF] [Spatial] [Dynamics] [Gene Regulation] Identifiable Feedback-Controlled Latent Flow for Unpaired Single-cell Spatio-Temporal Dynamics (NeurIPS 2026) [paper]

• [TraceSim] [Dynamics] [Benchmark] TraceSim: A Generative Simulator and Benchmark for Joint scRNA-seq and Lineage Tracing (NeurIPS 2026) [paper]

• [PACE] [Dynamics] PACE: Geometry-Aware Bridge Transport for Single-Cell Trajectory Inference (NeurIPS 2026) [paper] [preprint] [code]

• [TopoScape] [Spatial] [Morphology] [Multimodal] Layout Before Pixels: Topology-Anchored Transcriptome-to-Histology Generation (NeurIPS 2026) [paper]

• [GRNAgent] [Agent] [Gene Regulation] [Multimodal] GRNAgent: A Multimodal Graph Reasoning Agent for Gene Regulatory Network Inference (NeurIPS 2026) [paper]

• [AI in Single-Cell Genomics] [Review] [Foundation Model] [Virtual Cell] [Perturbation] [Gene Regulation] [Multimodal] Emerging frontiers for AI in single-cell genomics (Molecular Cell 2026) [paper]

• [bento-sc] [Foundation Model] [Evaluation & Measurement] [Benchmark] A systematic assessment of single-cell language model configurations (NAR Genomics and Bioinformatics 2026) [paper] [preprint] [code]

• [BulkFormer] [Foundation Model] [Representation Learning] [Related] BulkFormer: A large-scale foundation model for bulk transcriptomes (Cell Systems 2026) [paper] [preprint] [code]

• [CardamomOT] [Gene Regulation] [Dynamics] [Perturbation] CardamomOT: A mechanistic optimal transport-based framework for gene regulatory network inference, trajectory reconstruction and generative modeling (PLOS Computational Biology 2026) [paper] [preprint] [code] [documentation]

• [CauTrigger] [Perturbation] [Gene Regulation] [Intervention Design] CauTrigger: deciphering biological system state transitions by hierarchical causal decomposition (Nature Methods 2026) [paper] [code]

• [CCCvelo] [Dynamics] [Spatial] [Gene Regulation] Decoding cell state transitions driven by dynamic cell–cell communication in spatial transcriptomics (Nature Computational Science 2026) [paper] [code] [reproducibility]

• [3D Spheroid Cell Painting] [Morphology] [Perturbation] [Dataset] High-content morphological profiling by Cell Painting in 3D spheroids (Nature Communications 2026) [paper] [preprint] [code] [dataset]

• [Combinatorial TF Screening] [Perturbation] [Intervention Design] Pooled combinatorial screening identifies transcription factor sets that drive hematopoietic progenitor-like cell fate (Cell Systems 2026) [paper]

• [Counting Cells Baseline] [Evaluation & Measurement] [Morphology] [Benchmark] [Perturbation] Counting cells can accurately predict small-molecule bioactivity benchmarks (Nature Communications 2026) [paper] [preprint] [code] [data]

• [CoxFormer] [Representation Learning] [Spatial] [Multimodal] CoxFormer enables spatial omics inference with multimodal generative modeling (Nature Communications 2026) [paper] [preprint] [code] [documentation]

• [D-SPIN] [Gene Regulation] [Perturbation] D-SPIN constructs regulatory network models from scRNA-seq that reveal organizing principles of perturbation response (Cell 2026) [paper] [preprint] [code] [data]

• [DiffusionOT] [Dynamics] [Perturbation] [Gene Regulation] Learning stochastic dynamics and cell-fate landscapes from single-cell snapshots via optimal transport (Science Advances 2026) [paper] [code]

• [DynaVelo] [Dynamics] [Perturbation] [Gene Regulation] [Multimodal] [Intervention Design] Deep dynamical models of single-cell multiomic velocities predict loss-of-function and rescue perturbations in B cells (Nature Biotechnology 2026) [paper] [preprint] [code] [data]

• [ELISA] [Agent] [Multimodal] ELISA (Embedding-Linked Interactive Single-cell Agent): an interpretable hybrid generative Artificial Intelligence agent for expression-grounded discovery in single-cell genomics (Briefings in Bioinformatics 2026) [paper] [preprint] [code]

• [GLM-Prior] [Gene Regulation] [Foundation Model] [Multimodal] GLM-Prior: a genomic language model for transferable sequence-derived priors in gene regulatory network inference (Nature Communications 2026) [paper] [preprint] [code] [models]

• [GRNvelo] [Dynamics] [Gene Regulation] [Perturbation] Multiscale learning of gene network-driven phenotypic dynamics of single cells (Molecular Systems Biology 2026) [paper] [code]

• [LCD] [Perturbation] [Gene Regulation] Latent causal diffusions for single-cell perturbation modeling (PNAS 2026) [paper] [preprint] [code]

• [PaSCient] [Representation Learning] Learning multi-cellular representations of single-cell transcriptomics data enables characterization of patient-level disease states (Cell Systems 2026) [paper] [preprint]

• [RNA Fingerprinting] [Perturbation] [Representation Learning] Mapping transcriptional responses to cellular perturbations with RNA fingerprinting (Cell 2026) [paper] [preprint] [code] [data]

• [scDiffusion-X] [Multimodal] [Representation Learning] [Gene Regulation] [Perturbation] A multi-modal diffusion model with dual-cross-attention for multi-omics data generation and translation (Nature Communications 2026) [paper] [preprint] [code] [models]

• [scE2TM] [Representation Learning] [Perturbation] scE2TM improves single-cell embedding interpretability and reveals cellular perturbation signatures (Nature Communications 2026) [paper] [preprint] [code]

• [SCIGMA] [Spatial] [Multimodal] [Representation Learning] [Protein] [Tool] Scalable, generalizable and uncertainty-aware integration of spatial multiomics across diverse modalities and platforms with SCIGMA (Nature Genetics 2026) [paper] [preprint] [code] [documentation] [reproducibility]

• [Single-Cell Agent Benchmark] [Agent] [Benchmark] Benchmarking LLM-based agents for single-cell omics analysis (Genome Biology 2026) [paper] [preprint] [code] [dataset]

• [SPARCS] [Morphology] [Perturbation] [Tool] [Protein] [Multimodal] SPARCS enables scalable recovery of complex image-based phenotypes for genetic screening (Cell 2026) [paper] [preprint] [code] [image-processing code]

• [TFActProfiler] [Gene Regulation] [Perturbation] [Dataset] A transcription factor regulatory atlas for activity inference and perturbation prediction (Nucleic Acids Research 2026) [paper] [preprint] [code] [project] [dataset]

• [Whole-Cell Digital Twin] [Virtual Cell] [Morphology] [Dynamics] [Perturbation] Whole-cell particle-based digital twin simulations from 4D lattice light-sheet microscopy data (Cell 2026) [paper] [preprint]

• [ATLAS] [Multimodal] [Dynamics] [Representation Learning] ATLAS: Adaptive Temporal Learning for Single-Cell Multi-Omics Alignment and Dynamics (NeurIPS 2026) [paper]

• [Bigger Isn't Better] [Review] [Foundation Model] Bigger Isn’t Better: Why the Indiscriminate Scaling of Foundation Models Can’t Solve Biology (NeurIPS 2026) [paper]

• [Bulk Causal Recoverability] [Gene Regulation] [Evaluation & Measurement] On the Recoverability of Causal Relations from Bulk Gene Expression Data (NeurIPS 2026) [paper] [preprint]

• [Cellina] [Spatial] [Perturbation] [Representation Learning] Querying Counterfactuals on Tissue Graphs with Supervised Disentanglement (NeurIPS 2026) [paper] [preprint] [code] [documentation] [reproducibility]

• [ChanSFormer] [Morphology] [Representation Learning] ChanSFormer: A Channel Agnostic Vision Transformer for Multi-Channel Cell Painting Images (NeurIPS 2026) [paper]

• [Chem-PerturBridge] [Dataset] [Perturbation] [Benchmark] [Evaluation & Measurement] Chem-PerturBridge: a harmonized compendium of small molecule perturbation transcriptomic effects (NeurIPS 2026) [paper] [preprint] [code] [dataset] [analysis code] [LPM model code] [OP3 signatures code]

• [Cross-Cell-Line Controls] [Perturbation] [Evaluation & Measurement] [Benchmark] [Foundation Model] Cross-Cell-Line Perturbation Prediction Needs Controls (NeurIPS 2026) [paper]

• [CytoWave] [Perturbation] [Foundation Model] CytoWave: Perturbation-Centric Pretraining for Single-Cell Response Prediction (NeurIPS 2026) [paper]

• [DragH2ST] [Spatial] [Morphology] [Multimodal] [Gene Regulation] Dynamic Regulatory Graph Learning for Histology-to-Spatial Transcriptomics (NeurIPS 2026) [paper]

• [EVA-RNA] [Foundation Model] [Representation Learning] When Transcriptomic Foundation Models Scale: Domain-Focused Pretraining for Drug Development in Immunology and Inflammation (NeurIPS 2026) [paper] [models]

• [Flow Matching in ST] [Spatial] [Morphology] [Multimodal] When Does Flow Matching Help Deterministic Multimodal Prediction: A Spatial Transcriptomics Study (NeurIPS 2026) [paper] [code]

• [HoloGene] [Spatial] HoloGene: Learning to Lift Sliced Spatial Transcriptomics to Holistic 3D Gene Fields (NeurIPS 2026) [paper]

• [LeCellModel] [Representation Learning] [JEPA] LeCellModel: Interpretable Density Estimation over the Gene Expression Manifold (NeurIPS 2026) [paper]

• [MapPFN] [Perturbation] [Foundation Model] MapPFN: Learning Causal Perturbation Maps in Context (NeurIPS 2026) [paper] [preprint] [code] [project] [models] [dataset]

• [MorphGen] [Morphology] [Perturbation] MorphGen: Controllable Cell-Image Generation with Biological Representation Alignment (NeurIPS 2026) [paper] [preprint] [code]

• [MorphoHELM] [Benchmark] [Morphology] [Evaluation & Measurement] [Representation Learning] MorphoHELM: A Comprehensive Benchmark for Evaluating Representations for Microscopy-Based Morphology Assays (NeurIPS 2026) [paper] [preprint] [code]

• [MUSS] [Gene Regulation] [Multimodal] [Representation Learning] MUSS: A Multi-scale and Sequence-based Model for Single-cell Gene Regulation (NeurIPS 2026) [paper]

• [OrgBench / OrgFM] [Morphology] [Benchmark] [Foundation Model] [Representation Learning] Adapting Vision Transformers to Organoid Imaging (NeurIPS 2026) [paper]

• [PhenoAIR] [Morphology] [Agent] [Perturbation] From Retrieval to Reasoning: Agentic Mechanism Prediction from Cell Painting Profiles (NeurIPS 2026) [paper] [preprint] [code] [dataset]

• [Plausibility Is Not Prediction] [Evaluation & Measurement] [Perturbation] [Agent] [Virtual Cell] Plausibility Is Not Prediction: Contrastive Evidence for LLM-Based Cellular Perturbation Reasoning (NeurIPS 2026) [paper] [preprint]

• [scDecomp] [Multimodal] [Representation Learning] Decomposed Representations Mitigate the Alignment–Specificity Trade-off in Multi-Omics (NeurIPS 2026) [paper]

• [scPertCRL] [Perturbation] Causally Structured Differential Network Modeling for Single-Cell Perturbation Prediction (NeurIPS 2026) [paper]

• [scShapeBench] [Benchmark] [Dataset] [Evaluation & Measurement] scShapeBench: Discovering geometry from high dimensional scRNAseq data (NeurIPS 2026) [paper] [preprint] [dataset]

• [scTrilemma] [Representation Learning] scTrilemma: Balancing Identity, Invariance, and Fidelity in Single-Cell Representation Learning (NeurIPS 2026) [paper] [preprint] [code]

• [scWavelet] [Multimodal] [Representation Learning] Multi-Scale Representation Learning for Single-Cell Multi-Omics (NeurIPS 2026) [paper]

• [sMMC-22M] [Dataset] [Benchmark] [Spatial] [Morphology] [Multimodal] [Foundation Model] sMMC-22M: A Context-Aware Dataset and Benchmark for Single-Cell Spatial Transcriptomics (NeurIPS 2026) [paper]

• [spora] [Dataset] [Spatial] [Protein] [Multimodal] [Benchmark] [Tool] [Foundation Model] spora: A Unified Multimodal Dataset and Platform for Spatial Proteomics (NeurIPS 2026) [paper] [preprint] [code] [dataset] [documentation] [spora-io code]

• [Vermeer] [Protein] [Morphology] [Multimodal] Vermeer: Autoregressive generative modeling of microscopy predicts protein localization (NeurIPS 2026) [paper] [preprint] [code]

• [ALHENA] [Morphology] [Spatial] [Foundation Model] [Multimodal] [Representation Learning] [Tool] ALHENA, an H&E encoder distilled from transcriptomic foundation models for multi-scale slide analysis (bioRxiv 2026) [preprint] [code] [models] [project]

• [AmbiModBench] [Benchmark] [Perturbation] [Evaluation & Measurement] AmbiModBench: Benchmarking Gene Perturbation Prediction Beyond Shared Responses (arXiv 2026) [preprint] [code] [dataset]

• [AssayLoop] [Benchmark] [Intervention Design] [Perturbation] Biology-in-the-loop: Amortized Adaptive Hit Discovery in CRISPR Screens (arXiv 2026) [preprint] [code] [project] [models & data]

• [ATAC-to-RNA Benchmark] [Benchmark] [Multimodal] [Evaluation & Measurement] An autoencoder outperforms two transformers on prediction of gene expression from chromatin accessibility (arXiv 2026) [preprint] [code]

• [BatchBench] [Benchmark] [Morphology] [Evaluation & Measurement] BatchBench: Benchmarking AI Robustness to Batch Effects and Domain Shift in Biomedical Images - A JUMP Cell Painting Leaderboard (arXiv 2026) [preprint] [code] [leaderboard] [dataset]

• [BioBench] [Benchmark] [Evaluation & Measurement] [Representation Learning] Beyond Accuracy: Benchmarking the Reproducibility and Robustness of Single-Cell Embeddings (Research Square 2026) [preprint]

• [Biological Input Use] [Evaluation & Measurement] [Multimodal] [Foundation Model] [Agent] When Do Biological Reasoning Models Use Their Biological Inputs? (arXiv 2026) [preprint] [code] [project]

• [BioMod] [Foundation Model] [Representation Learning] [Perturbation] BioMod: Biological Module Tokenization for Single-Cell Foundation Modeling (bioRxiv 2026) [preprint]

• [Bison] [Perturbation] [Benchmark] Bison: Cross-Dataset Learning for Unseen-Compound Perturbation Prediction (arXiv 2026) [preprint]

• [Bolero] [Gene Regulation] [Multimodal] [Foundation Model] [Perturbation] Predicting cell-state-specific gene regulation from DNA across species (bioRxiv 2026) [preprint]

• [Brieflow-auto] [Agent] [Morphology] [Perturbation] [Tool] Harness-Driven Agentic Operation of Optical Pooled Screen Analysis (arXiv 2026) [preprint] [code] [Brieflow pipeline]

• [Calibrated Benchmarking Audit] [Evaluation & Measurement] [Perturbation] [Benchmark] Auditing Confidence-Calibrated Benchmarking of Single-Cell Perturbation Models (bioRxiv 2026) [preprint]

• [CEE] [Evaluation & Measurement] [Intervention Design] [Perturbation] [Benchmark] Cause-excluded Effect Evaluation: a framework for the faithful assessment of genetic perturbation target identification (bioRxiv 2026) [preprint]

• [CellAudit] [Evaluation & Measurement] [Agent] [Perturbation] [Virtual Cell] Discover, Falsify, Revise: Auditing Input-Use Claims from Source Code to Predictive Contribution in Agent-Discovered Cell Models (arXiv 2026) [preprint] [code] [project]

• [CellChain] [Benchmark] [Perturbation] [Evaluation & Measurement] [Agent] CellChain: Benchmarking Cross-Scale Evidence Use for Perturbation Reasoning (arXiv 2026) [preprint]

• [CellMAPP] [Morphology] [Perturbation] [Protein] [Tool] Archetype analysis of multiplexed antibody imaging reveals interpretable morphological programs in optical pooled screens (bioRxiv 2026) [preprint] [code]

• [CellNet] [Foundation Model] [Representation Learning] Panoramic modelling of single-cell transcriptomes at scale with CellNet (arXiv 2026) [preprint] [code] [project]

• [CELLO] [Spatial] [Morphology] [Multimodal] Towards Scalable Context-Aware Single-Cell Spatial Transcriptomics Prediction from Histology Images (NeurIPS 2026) [paper] [preprint] [code] [models] [dataset]

• [CellScribe] [Foundation Model] [Multimodal] [Representation Learning] From Closed-Set Annotation to In-Context Discovery: Biologically Grounded LLMs for Cellular Profiling (bioRxiv 2026) [preprint]

• [Centered Cross-Context Transfer] [Perturbation] Centered cross-context perturbation transfer on the public 2025 Virtual Cell Challenge benchmark (bioRxiv 2026) [preprint]

• [ChromaPert] [Perturbation] [Multimodal] [Gene Regulation] Beyond the Transcriptome: Chromatin-Informed Prediction of Cell-State-Dependent Perturbation Responses (arXiv 2026) [preprint]

• [CloneOT] [Perturbation] [Dynamics] CloneOT: predicting clone-specific perturbation responses with lineage-supervised neural inverse optimal transport (arXiv 2026) [preprint] [code] [documentation]

• [Concept Directions] [Foundation Model] [Representation Learning] [Related] Representing biological concepts as directions in foundation model embeddings (bioRxiv 2026) [preprint]

• [COSMIC] [Morphology] [Multimodal] [Foundation Model] [Spatial] Generative modeling reveals the connection between nuclear morphology and gene expression (bioRxiv 2026) [preprint] [code]

• [CP-BG-Bench] [Benchmark] [Morphology] [Evaluation & Measurement] [Representation Learning] Same Encoder, Different Winner: A Paired-View Framework for Cell Painting Encoder Evaluation (arXiv 2026) [preprint]

• [CRISPR Escape Dilution] [Evaluation & Measurement] [Perturbation] Escaping cells dilute CRISPR screen responses unequally across cell types (arXiv 2026) [preprint] [code] [software archive]

• [CUBE] [Morphology] [Spatial] [Protein] [Multimodal] [Representation Learning] CUBE: Multimodal Representation Learning Reveals Biological Structure Across Histomorphology, Spatial Protein Phenotypes, and Transcriptome-Associated Signals (bioRxiv 2026) [preprint]

• [DE Evaluation Pitfall] [Evaluation & Measurement] [Perturbation] [Benchmark] A Pitfall in the Evaluation of Gene Perturbation Models (arXiv 2026) [preprint]

• [DeepMAST] [Agent] [Spatial] [Tool] An evidence-governed agent harness for spatial transcriptomics prioritizes ICAM-1 as a target in lung fibrosis (bioRxiv 2026) [preprint]

• [DELPHAI] [Perturbation] DELPHAI predicts heterogeneous perturbation responses with learned single-cell fitness (bioRxiv 2026) [preprint]

• [FM Perturbation Utility] [Foundation Model] [Perturbation] [Evaluation & Measurement] [Benchmark] [Representation Learning] Foundation Models Improve Perturbation Response Prediction (bioRxiv 2026) [preprint] [code]

• [GEM-GPT] [Intervention Design] [Perturbation] [Multimodal] [Foundation Model] GEM-GPT Enables Personalized Cell Type-Resolved Therapeutic Design for Systems Pharmacology (bioRxiv 2026) [preprint]

• [GEMOT] [World Model] [Agent] [Dynamics] [Related] GEMOT: Towards Mechanistic World Models for Biology (bioRxiv 2026) [preprint]

• [GENIE] [Agent] [Morphology] [Tool] GENIE: a multi-agent system for closed-loop autonomous microscopy (bioRxiv 2026) [preprint]

• [GENOAR] [Dataset] [Agent] [Tool] GENOAR: Global Engine for Navigating the Omnicell space via Agent-based Research (bioRxiv 2026) [preprint]

• [GeoSinkhorn Flow] [Perturbation] [Dynamics] [Protein] GeoSinkhorn Flow: Geometry-Aware Flow Matching for Perturbation-Conditioned Dynamics in Single cell Data Phenoscapes (bioRxiv 2026) [preprint]

• [hallsim] [Virtual Cell] [Dynamics] [Perturbation] [Tool] An AI-ready compositional framework for mechanistic aging research and in silico intervention testing (bioRxiv 2026) [preprint] [code]

• [H&E-to-ST Error Structure] [Evaluation & Measurement] [Spatial] [Morphology] [Multimodal] Is H&E Image-to-Spatial Transcriptomics Simpler Than It Looks? (arXiv 2026) [preprint]

• [HistAgent] [Agent] [Foundation Model] [Spatial] [Morphology] [Multimodal] HistAgent: an evidence-grounded agent for spatial molecular reasoning from routine histology (Research Square 2026) [preprint] [code (project)] [models]

• [Immune World Model] [World Model] [Agent] [Perturbation] [Intervention Design] An immune world model for multiscale forecasting and therapeutic hypothesis generation (arXiv 2026) [preprint]

• [InverseBench] [Intervention Design] [Benchmark] [Evaluation & Measurement] [Perturbation] Information contracts determine what inverse perturbation benchmarks measure (bioRxiv 2026) [preprint]

• [Mechanisms Matter] [Perturbation] [Evaluation & Measurement] [Benchmark] Mechanisms Matter: Transportability of Cellular Perturbation Effects (bioRxiv 2026) [preprint] [code]

• [MILK] [Tool] [Representation Learning] Global tree encoding of atlas-scale single-cell genomics (bioRxiv 2026) [preprint]

• [MRFlow] [Virtual Cell] [Perturbation] MRFlow: Mechanism-Structured Residual Flow Matching for Virtual Cell Perturbation Prediction (arXiv 2026) [preprint]

• [OmicsBench] [Agent] [Benchmark] [Evaluation & Measurement] OmicsBench: a long-horizon multi-turn benchmark of AI agents on research-study-scale bioinformatics (arXiv 2026) [preprint] [code] [data]

• [OmniVCBench] [Virtual Cell] [Benchmark] [Agent] [Multimodal] [Evaluation & Measurement] OmniVCBench: Benchmarking Evidence-Grounded Multimodal Reasoning Towards AI Virtual Cells (arXiv 2026) [preprint] [code]

• [P2P] [Perturbation] [Virtual Cell] P2P: Cross-View Population Denoising for Unpaired Single-Cell Perturbation Response Prediction (arXiv 2026) [preprint]

• [EL-PERTURB] [Benchmark] [Perturbation] [Evaluation & Measurement] Genetic Perturbation Prediction Models Fail Out-of-Distribution: A Rigorous Benchmark and Systematic Evaluation (bioRxiv 2026) [preprint] [code (project)]

• [Perturbation-Strength Averaging] [Evaluation & Measurement] [Perturbation] Averaging in gene perturbation benchmarks hides perturbation-strength variation: Consequences for model evaluation (bioRxiv 2026) [preprint]

• [PerturbBridge (image)] [Perturbation] [Morphology] [Virtual Cell] PerturbBridge: Direct Phenotypic Supervision for Cellular Perturbation Generation (arXiv 2026) [preprint] [code]

• [PerturbEnergy] [Perturbation] [Representation Learning] Learning Perturbation-Specific Representations through Conditional Energy Prior for Cellular Response Prediction (arXiv 2026) [preprint] [code]

• [PETRO] [Perturbation] [Evaluation & Measurement] [Tool] PETRO enables robust estimation of on-target perturbation effects in Perturb-seq (bioRxiv 2026) [preprint] [code (project)]

• [Pooled-Screen Pseudoreplication] [Evaluation & Measurement] [Perturbation] Shared clones make random cell halves look reproducible in pooled screens (arXiv 2026) [preprint] [code] [software archive]

• [PRAXIS-VirtualCell] [Virtual Cell] [Agent] [Tool] [Perturbation] PRAXIS-VirtualCell: A Programmable and Trustworthy Framework for Agentic Virtual Cell Experiments (arXiv 2026) [preprint]

• [Reliability Transfer] [Foundation Model] [Evaluation & Measurement] [Perturbation] Prediction transfer does not imply reliability transfer in single-cell foundation models (bioRxiv 2026) [preprint]

• [RevPert] [Intervention Design] [Perturbation] RevPert: predicting candidate drivers of transcriptomic state transitions via gallery-native reverse perturbation (bioRxiv 2026) [preprint]

• [scDrug Response Benchmark] [Benchmark] [Perturbation] [Evaluation & Measurement] Predicting Pre-treatment Resistance or Post-treatment Effect? A Systematic Benchmarking of Single-Cell Drug Response Models (bioRxiv 2026) [preprint]

• [SC-islet Digital Model] [Intervention Design] [Gene Regulation] [Dynamics] [Multimodal] [Perturbation] Predictive control of human pancreatic cell fate using a digital model of in vitro differentiation (bioRxiv 2026) [preprint]

• [scACORN] [Agent] [Foundation Model] [Multimodal] scACORN: Context-engineered agent orchestration of specialized small language models for single-cell transcriptomic interpretation (bioRxiv 2026) [preprint]

• [scDEFT] [Perturbation] scDEFT: A deep learning framework for drug-effect prediction and counterfactual reasoning (arXiv 2026) [preprint]

• [scFM Confidence Reliability] [Foundation Model] [Evaluation & Measurement] Cross-domain confidence reliability and remappability of frozen single-cell representations (bioRxiv 2026) [preprint]

• [scFM Layer Probes] [Foundation Model] [Evaluation & Measurement] [Benchmark] [Perturbation] [Representation Learning] Beyond the Final Layer: Evaluating Single-Cell Foundation Model Representations for Molecular Prediction (bioRxiv 2026) [preprint]

• [scFM Long-Tail Benchmark] [Benchmark] [Foundation Model] [Evaluation & Measurement] Rethinking Class Imbalance for Single-Cell Foundation Models: A Systematic Benchmark Across Architectures and Long-Tail Loss Functions (arXiv 2026) [preprint] [code]

• [scFM Low-Data Benchmark] [Foundation Model] [Benchmark] [Evaluation & Measurement] [Representation Learning] Benchmarking single-cell foundation models reveals an advantage in the low-data regime for cell-type and disease-state prediction (bioRxiv 2026) [preprint]

• [SHERLOCK] [Perturbation] [Representation Learning] SHERLOCK: Structured representation learning and causal inference of downstream perturbation effects (bioRxiv 2026) [preprint] [code]

• [Signature Recontextualization] [Perturbation] [Benchmark] [Evaluation & Measurement] [Tool] A benchmarking framework for perturbational signature recontextualization across biological contexts (bioRxiv 2026) [preprint] [code (project)]

• [Single-Cell Split Leakage] [Evaluation & Measurement] [Benchmark] [Tool] Cell-level random splits leak group-owned answers in single-cell benchmarks (bioRxiv 2026) [preprint]

• [Slide-Level Batch Structure] [Evaluation & Measurement] [Spatial] [Foundation Model] [Morphology] [Multimodal] [Representation Learning] Slide-level batch structure limits histology-guided supervision of transcriptomic foundation models (bioRxiv 2026) [preprint] [code] [models]

• [spaGFM] [Foundation Model] [Spatial] [Representation Learning] [Perturbation] spaGFM is a scalable graph foundation model for spatial transcriptomics analyses (Research Square 2026) [preprint] [code] [reproducibility]

• [SpatialOmicsLab] [Agent] [Spatial] [Tool] SpatialOmicsLab: an integrated research environment for AI co-scientists in spatial transcriptomics (bioRxiv 2026) [preprint] [code]

• [SafeDelta] [Evaluation & Measurement] [Perturbation] A split-control criterion for few-shot adaptation in single-cell perturbation prediction (bioRxiv 2026) [preprint] [code]

• [Beyond the Global Trend] [Evaluation & Measurement] [Perturbation] Transcriptional perturbation effects remain structured beyond the leading global trend (bioRxiv 2026) [preprint] [code] [data]

• [SubcellPert] [Benchmark] [Perturbation] [Spatial] [Evaluation & Measurement] [Virtual Cell] SubcellPert: Benchmarking Subcellular Perturbation Prediction (bioRxiv 2026) [preprint] [code] [dataset] [project]

• [SVC] [Spatial] [Representation Learning] [Perturbation] SVC: A Vision Transformer-based Spatially and Virtually Embedded Cell Model for Deciphering Subcellular Spatial Transcriptomic Heterogeneity (bioRxiv 2026) [preprint] [code] [documentation] [reproducibility]

• [SynGeneFlow] [Perturbation] Synergy-Aware Residual Flow Matching for Single-Cell Perturbation Prediction (arXiv 2026) [preprint]

• [Tahoe-100M Response Axes] [Evaluation & Measurement] [Perturbation] Two largely independent axes of drug transcriptional response in Tahoe-100M (Research Square 2026) [preprint] [code]

• [TDEP] [Perturbation] TDEP: Predicting Target Perturbation Response from Drug-Induced Transcriptomes (Research Square 2026) [preprint]

• [TempoBridge] [Perturbation] [Dynamics] TempoBridge: Source-Conditioned Flow Matching with Optimal Transport Couplings for Single-Cell Population Transitions (arXiv 2026) [preprint]

• [Training Scale Re-evaluation] [Foundation Model] [Evaluation & Measurement] [Representation Learning] Re-evaluating the benefit of training scale in cell foundation models (bioRxiv 2026) [preprint]

• [UOT-IOT] [Dynamics] Identifiable inverse optimal transport for single-cell state transitions (bioRxiv 2026) [preprint]

• [VCLMU] [Virtual Cell] [World Model] [Perturbation] VCLMU: Mechanism-Centric Virtual Cell World Modeling for Perturbation Response (arXiv 2026) [preprint]

• [Virtual ST Gene Representations] [Spatial] [Evaluation & Measurement] [Morphology] [Foundation Model] [Representation Learning] [Multimodal] Pretrained gene representations transfer mean expression more broadly than spatial patterns in virtual spatial transcriptomics (bioRxiv 2026) [preprint] [code]

• [Virtual-ST Uncertainty] [Evaluation & Measurement] [Benchmark] [Spatial] [Morphology] Probabilistic uncertainty in virtual spatial transcriptomics from histology is largely redundant with the predicted mean (bioRxiv 2026) [preprint] [code]

• [scPerturbBench (Sci Adv)] [Benchmark] [Perturbation] [Evaluation & Measurement] A systematic comparison of single-cell perturbation response prediction models (Science Advances 2026) [paper] [preprint] [code]

• [PrePR-CT] [Perturbation] Predicting and interpreting cell-type-specific drug responses in the small-data regime using inductive priors (Nature Machine Intelligence 2026) [paper] [preprint] [code]

• [scPILOT] [Perturbation] Predicting Single-Cell Perturbation Responses Across Biological Contexts With a Deep Generative Model Integrating Optimal Transport (Advanced Science 2026) [paper] [code]

• [PRIM] [Perturbation] Incorporating valuable prior knowledge to improve deep learning prediction of genetic perturbation responses (Genome Research 2026) [paper]

• [scDEBART] [Foundation Model] [Perturbation] scDEBART: Predicting in silico Single-Cell Perturbation Responses via Large-Scale Differential Expression Learning (ICML 2026) [paper]

• [AdaPert] [Perturbation] Learning Adaptive Perturbation-Conditioned Contexts for Robust Transcriptional Response Prediction (ICML 2026) [paper] [preprint]

• [CURE] [Intervention Design] [Perturbation] [Multimodal] Reading the Cell, Designing the Cure: Perturbation-Conditioned Molecular Diffusion for Function-Oriented Drug Design (ICML 2026) [paper] [code]

• [AROMA] [Virtual Cell] [Perturbation] [Multimodal] AROMA: Augmented Reasoning Over a Multimodal Architecture for Virtual Cell Genetic Perturbation Modeling (ACL Findings 2026) [preprint] [code]

• [MultiFlow] [Perturbation] [Multimodal] MultiFlow: coupled flow matching for predicting single-cell multiomic perturbation responses in unseen cellular contexts (bioRxiv 2026) [preprint] [code (project)]

• [DRIFT] [Perturbation] [Representation Learning] DRIFT: Disentangled Responsive-Invariant Flow Transport for Single-Cell Perturbation Prediction (arXiv 2026) [preprint]

• [scEGFlow] [Perturbation] [Morphology] Generalizable single-cell perturbation response prediction using energy-guided flow matching (arXiv 2026) [preprint]

• [CIFM] [Virtual Cell] [Spatial] [Foundation Model] [Perturbation] [Intervention Design] A Generative Virtual Tissue Model Enables Computational Design of Therapeutic Perturbation Strategies (bioRxiv 2026) [preprint]

• [PRiMeFlow] [Perturbation] PRiMeFlow: Capturing Complex Expression Heterogeneity in Perturbation Response Modelling (arXiv 2026) [preprint]

• [scDrugPerturb-Bench] [Benchmark] [Evaluation & Measurement] [Perturbation] A mechanism-annotated benchmark reveals limited fidelity to drug-response signatures in single-cell perturbation models (bioRxiv 2026) [preprint]

• [Intervention Geometry] [Evaluation & Measurement] [Perturbation] [Virtual Cell] Virtual-cell models compress unseen intervention geometry through a target-specific generalization bottleneck (bioRxiv 2026) [preprint]

• [Calibrated Metrics] [Evaluation & Measurement] [Benchmark] [Perturbation] Deep learning perturbation models can outperform baselines on calibrated metrics (Nature Biotechnology 2026) [paper]

• [Schmidt Center T-cell Challenge] [Benchmark] [Perturbation] [Intervention Design] A community machine learning challenge to predict the effects of gene perturbations on T cell differentiation for cancer immunotherapy (bioRxiv 2026) [preprint] [challenge]

• [AlphaGenome] [Foundation Model] [Gene Regulation] [Multimodal] [Related] Advancing regulatory variant effect prediction with AlphaGenome (Nature 2026) [paper] [preprint] [code (project)]

• [ChromBERT] [Foundation Model] [Gene Regulation] [Representation Learning] ChromBERT: A foundation model for learning interpretable representations for context-specific transcriptional regulatory networks (Cell Genomics 2026) [paper] [preprint] [code]

• [MaxToki] [Foundation Model] [Dynamics] [Perturbation] [Intervention Design] Temporal AI model predicts drivers of cell state trajectories across human aging (bioRxiv 2026) [preprint] [code]

• [CellMSA] [Foundation Model] [Representation Learning] CellMSA: Context Modeling for Single-Cell Representation Learning (NeurIPS 2026) [paper] [preprint] [code]

• [scELMo] [Foundation Model] [Representation Learning] [Multimodal] Embeddings from language models are good learners for single-cell data analysis (Patterns 2026) [paper] [preprint] [code]

• [SQUALL] [Foundation Model] [Spatial] [Multimodal] [Morphology] A multimodal foundation model to integrate histology with spatial molecular programs (bioRxiv 2026) [preprint]

• [OmicsFM] [Foundation Model] [Protein] [Representation Learning] OmicsFM brings proteomics into the foundation model era (bioRxiv 2026) [preprint]

• [Cross-Modal Proteomics Pretraining] [Foundation Model] [Protein] [Multimodal] [Representation Learning] Single-cell foundation models benefit from cross-modal training: adding proteomics data beats parameter scaling (bioRxiv 2026) [preprint]

• [BioMicroscopyProfiler] [Morphology] [Multimodal] [Representation Learning] [Perturbation] Intervention-Aware Multiscale Representation Learning from Imaging Phenomics and Perturbation Transcriptomics (CVPR 2026) [paper] [preprint] [code (project)]

• [TxFM] [Foundation Model] [Representation Learning] [Evaluation & Measurement] Effective Biological Representation Learning by Masking Gene Expression (NeurIPS 2026) [paper] [preprint] [code]

• [AssayBench] [Agent] [Benchmark] [Virtual Cell] [Evaluation & Measurement] AssayBench: An Assay-Level Virtual Cell Benchmark for LLMs and Agents (NeurIPS 2026) [paper] [preprint] [code] [dataset]

• [iPSC CRISPRi Atlas] [Dataset] [Perturbation] [Gene Regulation] A genome-scale CRISPRi perturbation atlas of human induced pluripotent stem cells (Nature Biotechnology 2026) [paper] [preprint] [code] [dataset]

• [In Vivo Brain Perturb-seq] [Dataset] [Perturbation] Genome-scale functional mapping of the mammalian whole brain with in vivo Perturb-seq (bioRxiv 2026) [preprint]

• [SPAC-seq] [Dataset] [Spatial] [Perturbation] Uncovering spatially resolved functional genomics with CRISPR screen sequencing (Cell 2026) [paper]

• [4D Minimal Cell] [Virtual Cell] [Dynamics] Bringing the genetically minimal cell to life on a computer in 4D (Cell 2026) [paper]

• [CellAgent] [Agent] [Tool] CellAgent: LLM-Driven Multi-Agent Framework for Natural Language-Based Single-Cell Analysis (ICLR 2026) [paper] [preprint]

• [LabCompass] [Intervention Design] [Perturbation] [Tool] Generative inverse design steers human hematopoietic cell fate in vitro (bioRxiv 2026) [preprint] [code] [reproducibility] [data & checkpoints]

• [PIE] [Perturbation] [Multimodal] PIE: Generalizing perturbation effects across unseen perturbations, contexts and datasets (bioRxiv 2026) [preprint] [code] [data & knowledge sources]

• [Recoverable Resolution] [Evaluation & Measurement] [Benchmark] [Perturbation] [Virtual Cell] The recoverable resolution of cellular perturbation-response prediction (bioRxiv 2026) [preprint] [code] [source data]

• [Pop-Corn] [Perturbation] [Spatial] [Intervention Design] Pop-Corn: Predicting Perturbation Phenotype Effects Across Single-Cell and Spatial Contexts (bioRxiv 2026) [preprint]

• [NexuST] [Foundation Model] [Spatial] [Representation Learning] NexuST: A Hierarchical Foundation Model for Spatial Transcriptomics (bioRxiv 2026) [preprint]

• [EpiZoo] [Foundation Model] [Gene Regulation] [Multimodal] [Representation Learning] EpiZoo: a DNA sequence-aware foundation model for cross-species single-cell epigenomics (bioRxiv 2026) [preprint]

• [PerturbBridge] [Perturbation] [Dynamics] [Representation Learning] PerturbBridge: Conditional Latent Schrödinger Bridge for Single-Cell Perturbation Response Prediction (bioRxiv 2026) [preprint]

• [BioPert] [Perturbation] [Representation Learning] A biological-response compound representation allows chemical perturbation prediction across cell lines (bioRxiv 2026) [preprint]

• [SPECTRA] [Perturbation] [Gene Regulation] SPECTRA: predicting cellular perturbation responses with Graph Learning over Gene Regulatory Networks (bioRxiv 2026) [preprint]

• [ISP³ Platform] [Tool] [Foundation Model] [Perturbation] [Intervention Design] [Dynamics] ISP³ Platform powered by Geneformer: Framework for Cross-Species, Sequential, and Multi-Gene In Silico Perturbation Screens with Application to iPS Cell State Transitions (bioRxiv 2026) [preprint]

• [Aging scFM Benchmark] [Benchmark] [Foundation Model] [Evaluation & Measurement] [Related] Benchmarking single-cell foundation models for aging biology (bioRxiv 2026) [preprint]

• [CellRFT] [Perturbation] [Evaluation & Measurement] CellRFT: Reinforcement Fine-Tuning for Single-Cell Perturbation Modeling (arXiv 2026) [preprint]

• [PopPert] [Perturbation] PopPert: Population-level Joint-Distribution Modeling for Single-Cell Perturbation Prediction (arXiv 2026) [preprint] [code]

• [STP-BENCH] [Benchmark] [Spatial] [Morphology] [Evaluation & Measurement] STP-BENCH: A Unified Systematic Benchmark for Virtual Spatial Transcriptomics from Histopathology Images (arXiv 2026) [preprint] [code]

• [RAGCell] [Foundation Model] [Multimodal] [Representation Learning] RAGCell: Retrieval-Augmented Generation as Supervision for Versatile Single-cell Analysis (Bioinformatics 2026) [paper] [preprint]

• [stFormer] [Foundation Model] [Spatial] [Gene Regulation] [Representation Learning] stFormer integrates spatial ligand signaling into a foundation model for spatial transcriptomics (Cell Reports Methods 2026) [paper]

• [scLDM] [Perturbation] [Dynamics] scLDM: a conditional diffusion framework for single-cell perturbation prediction (Bioinformatics 2026) [paper] [code]

• [CSGDA] [Perturbation] [Representation Learning] CSGDA: A Cell State-Guided Graph Domain Adaptation Network for Single-Cell Drug Response Prediction (Bioinformatics 2026) [paper] [preprint]

• [DePass] [Tool] [Spatial] [Multimodal] [Representation Learning] The dual-enhanced graph learning framework DePass allows paired data integration in single-cell and spatial multiomics (Nature Cell Biology 2026) [paper] [code] [documentation]

• [IRIS] [Perturbation] [Dynamics] [Intervention Design] [Dataset] Reconstructing signaling histories of single cells via perturbation screens and transfer learning (Nature Methods 2026) [paper] [dataset]

• [MAP] [Perturbation] [Foundation Model] [Multimodal] [Representation Learning] A knowledge-driven framework for predicting single-cell responses for unprofiled drugs (Nature Machine Intelligence 2026) [paper] [preprint] [code] [dataset]

• [SCMBench] [Benchmark] [Foundation Model] [Multimodal] [Evaluation & Measurement] SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration (Nature Communications 2026) [paper] [code]

• [GeneformerV2] [Foundation Model] [Representation Learning] [Gene Regulation] [Perturbation] [Tool] Scaling and quantization of large-scale foundation model enables resource-efficient predictions in network biology (Nature Computational Science 2026) [paper] [code] [dataset] [documentation]

• [XPert] [Perturbation] [Dynamics] [Gene Regulation] Modelling drug-induced cellular perturbation responses with a biologically informed dual-branch transformer (Nature Machine Intelligence 2026) [paper] [code] [dataset]

• [scDMC] [Foundation Model] [Representation Learning] [Gene Regulation] scDMC: Unlocking biological insight from single-cell data with an interpretable dual-stream foundation model (Genome Biology 2026) [paper]

• [CellVQ] [Foundation Model] [Representation Learning] CellVQ: Illuminating cell states by a comprehensive and interpretable single cell foundation model (Nature Communications 2026) [paper]

• [SpatialFormer] [Foundation Model] [Spatial] [Multimodal] [Representation Learning] SpatialFormer: universal spatial representation learning from subcellular molecular to multicellular landscapes (Nature Computational Science 2026) [paper]

• [VirTues] [Virtual Cell] [Foundation Model] [Spatial] [Protein] [Representation Learning] The Virtual Tissues foundation model resolves spatial proteomics across scales (Nature 2026) [paper]

• [HEX] [Spatial] [Morphology] [Protein] [Multimodal] AI-enabled virtual spatial proteomics from histopathology for interpretable biomarker discovery in lung cancer (Nature Medicine 2026) [paper]

• [spEMO] [Foundation Model] [Spatial] [Morphology] [Multimodal] [Representation Learning] Leveraging Multi-Modal Foundation Models for Analyzing Spatial Multi-Omic and Histopathology Data (Nature Biomedical Engineering 2026) [paper]

• [Scaling Is Much Pain] [Foundation Model] [Benchmark] [Evaluation & Measurement] Scaling up training dataset size for transcriptomic AI models is much pain with little gain (Nature Methods 2026) [paper]

• [Biomedical FM Benchmark] [Foundation Model] [Benchmark] [Evaluation & Measurement] [Related] Benchmarking biomedical foundation models (Nature Methods 2026) [paper]

• [scTranslation] [Benchmark] [Multimodal] [Evaluation & Measurement] scTranslation: A Comprehensive Benchmark for Single-Cell Multi-Omics Modality Translation (KDD 2026) [paper]

• [Interpretation, Extrapolation & Perturbation] [Review] [Foundation Model] [Perturbation] [Gene Regulation] Interpretation, extrapolation and perturbation of single cells (Nature Reviews Genetics 2026) [paper]

• [AI Digital Organism] [Review] [Virtual Cell] [World Model] [Related] How to build an AI-driven digital organism (Nature Medicine 2026) [paper]

• [World Models for Biomedicine] [World Model] [Review] [Related] World models for biomedicine (Cell 2026) [paper]

• [Fifteen Challenges] [Review] [Virtual Cell] [Related] Fifteen challenges for generative AI applications to cell biology (Cell 2026) [paper]

• [Compositional Foundation Models] [Review] [Foundation Model] [Multimodal] From modality-specific to compositional foundation models for cell biology (Cell Systems 2026) [paper]

• [Nuisance Robustness] [Foundation Model] [Benchmark] [Evaluation & Measurement] Robustness to nuisance perturbations enables unsupervised evaluation of single-cell foundation models (bioRxiv 2026) [preprint]

• [Scaling Recipes] [Foundation Model] [Benchmark] [Evaluation & Measurement] Scaling recipes for single-cell RNA sequencing foundation models: when do scaling laws hold? (bioRxiv 2026) [preprint]

• [Accessible scFM Deployment] [Foundation Model] [Evaluation & Measurement] [Tool] Accessible and reproducible deployment reveals the practical boundaries of single-cell foundation models (bioRxiv 2026) [preprint]

• [Parameter-Free Representations] [Foundation Model] [Benchmark] [Evaluation & Measurement] [Representation Learning] [Related] Parameter-free representations outperform single-cell foundation models on downstream benchmarks (arXiv 2026) [preprint]

• [PertResolve] [Evaluation & Measurement] [Benchmark] [Perturbation] [Virtual Cell] [Tool] Measurement resolution constrains fine-grained perturbation prediction (Manuscript 2026) [preprint] [code] [project] [dataset]

• [Signal, Bounds & Baselines] [Evaluation & Measurement] [Benchmark] [Perturbation] [Virtual Cell] Signal, Bounds, and Baselines: Principles for Evaluating Virtual Cell Perturbation Models (bioRxiv 2026) [preprint] [code]

• [Metric Failure Modes] [Evaluation & Measurement] [Benchmark] [Perturbation] [Virtual Cell] Evaluating Single-Cell Perturbation Response Models Is Far from Straightforward (bioRxiv 2026) [preprint]

• [VCBench (In-the-Wild)] [Benchmark] [Evaluation & Measurement] [Perturbation] [Virtual Cell] Benchmarking virtual cell models for in-the-wild perturbation response (arXiv 2026) [preprint] [code] [project]

• [VCDesign] [Intervention Design] [Perturbation] [Virtual Cell] VCDesign: Finite-Budget Intervention Design for Virtual Cells (Manuscript 2026) [preprint] [code] [project] [dataset]

• [NUDGE] [Intervention Design] [Gene Regulation] [Dynamics] [Related] Uncovering minimal control of cell fate by natural dynamics (PNAS 2026) [paper] [code]

• [Speciesformer] [Virtual Cell] [Foundation Model] [Multimodal] [Perturbation] [Representation Learning] Speciesformer learns conserved cellular states for cross-species generative virtual cell modeling (bioRxiv 2026) [preprint]

• [DeepSCENIC] [Gene Regulation] [Multimodal] [Perturbation] DeepSCENIC: transfer learning from sequence-to-function models enables causal gene regulatory network inference (bioRxiv 2026) [preprint] [code]

• [scKITE] [Foundation Model] [Representation Learning] Towards a knowledge-enhanced single-cell foundation model (arXiv 2026) [preprint] [code]

• [PHAROS] [Perturbation] [Intervention Design] [Virtual Cell] PHAROS: turning single-cell perturbation models into target-directed drug-combination screens (bioRxiv 2026) [preprint] [code] [reproduce]

• [ProteinTalks] [Virtual Cell] [Foundation Model] [Perturbation] [Protein] [Dynamics] [Intervention Design] An operational perturbation proteomics-based virtual cell model (Nature 2026) [paper] [preprint] [code] [中文解读] [ask deepwiki]

• [Virtual Cell Challenge 2026] [Benchmark] [Perturbation] Virtual Cell Challenge 2026: Benchmarking zero-shot generalization across cellular contexts (Cell 2026) [paper] [challenge]

• [LucaCell] [Foundation Model] [Representation Learning] [Multimodal] LucaCell: a sequence-centric foundation model for cross-species single-cell analysis (bioRxiv 2026) [preprint] [code]

• [AnnFlux] [Perturbation] [Dynamics] AnnFlux: object-conditioned neural stochastic differential equations for single-cell perturbation dynamics (bioRxiv 2026) [preprint]

• [scRep] [Foundation Model] [Representation Learning] scRep: A Latent-Space Self-Distilled Foundation Model for Single-Cell Representation Learning (bioRxiv 2026) [preprint]

• [Cell-o1] [Agent] [Benchmark] Cell-o1: Training LLMs to Solve Single-Cell Reasoning Puzzles with Reinforcement Learning (Bioinformatics 2026) [paper] [preprint] [code] [hugging face] [ask deepwiki]

• [RegVelo] [Dynamics] [Gene Regulation] [Perturbation] [Multimodal] RegVelo: Gene-Regulatory-Informed Dynamics of Single Cells (Cell 2026) [paper] [preprint] [code] [ask deepwiki]

• [TranscriptFormer] [Foundation Model] [Representation Learning] TranscriptFormer: A Generative Cell Atlas across 1.5 Billion Years of Evolution (Science 2026) [paper] [preprint] [code] [ask deepwiki]

• [CellAtria] [Agent] [Tool] An Agentic AI Framework for Ingestion and Standardization of Single-Cell RNA-Seq Data Analysis (npj Artificial Intelligence 2026) [paper] [preprint] [code] [中文解读] [ask deepwiki]

• [CellVoyager] [Agent] [Benchmark] CellVoyager: AI CompBio Agent Generates New Insights by Autonomously Analyzing Biological Data (Nature Methods 2026) [paper] [preprint] [code] [中文解读] [ask deepwiki]

• [Biomni] [Agent] [Related] Autonomous Biomedical Research with an Artificial Intelligence Agent (Science 2026) [paper] [preprint] [code] [ask deepwiki]

• [STATE] [Perturbation] [Foundation Model] [Benchmark] Predicting cellular responses to perturbation across diverse contexts with State (Cell 2026) [paper] [preprint] [code] [ask deepwiki]

• [UniPert-G2CP] [Perturbation] [Multimodal] UniPert-G2CP Bridges Genetic and Chemical Screens from Molecular Representation to Phenotype Modeling (Cell 2026) [paper] [preprint] [code] [ask deepwiki]

• [Cell Shapes] [Morphology] [Protein] [Perturbation] Cell shapes decode molecular phenotypes in image-based spatial proteomics (Cell Systems 2026) [paper] [preprint] [code]

• [scLong] [Foundation Model] [Representation Learning] [Perturbation] [Gene Regulation] scLong: A Billion-Parameter Foundation Model for Capturing Long-Range Gene Context in Single-Cell Transcriptomics (Nature Communications 2026) [paper] [preprint] [code] [ask deepwiki]

• [UCE] [Foundation Model] [Representation Learning] Universal Cell Embedding Provides a Foundation Model for Cell Biology (Nature 2026) [paper] [preprint] [code] [ask deepwiki]

• [CellOS] [Virtual Cell] [World Model] [JEPA] [Foundation Model] [Representation Learning] [Multimodal] [Perturbation] CellOS: Learning a World Model of Cellular State through Joint Embedding Prediction (bioRxiv 2026) [preprint]

• [Chreode] [World Model] [Foundation Model] [Dynamics] [Perturbation] Chreode: A Cell World Model for One-Step Temporal Dynamics and Perturbation Prediction (arXiv 2026) [preprint] [code]

• [VCWM] [Virtual Cell] [World Model] [Review] A world model of the virtual cell (Cell 2026) [paper] [technical report]

• [World Model Gaps] [World Model] [Virtual Cell] [Benchmark] [Evaluation & Measurement] What Makes a Virtual Cell a World Model? Three Gaps, Three Experiments, and a Roadmap (Research Square 2026) [preprint]

• [Biomedical World Models] [World Model] [Review] [Related] Towards World Models in Biomedical Research (arXiv 2026) [preprint]

• [CENO] [World Model] [Foundation Model] [Gene Regulation] [Intervention Design] [Related] CENO: A Genome-Scale World Model for Evolutionary Sequence Interpretation and Programmable Regulatory Design (bioRxiv 2026) [preprint] [code]

• [Cell-JEPA] [JEPA] [Foundation Model] [Representation Learning] [Perturbation] Cell-JEPA: Latent Representation Learning for Single-Cell Transcriptomics (arXiv 2026) [preprint]

• [SpatialJEPA] [JEPA] [Spatial] [Multimodal] [Representation Learning] SpatialJEPA: JEPA-Inspired Graph-Context Distillation for Spatially Aware Multiomics Integration (bioRxiv 2026) [preprint] [code]

• [BioM-JEPA] [JEPA] [Foundation Model] [Representation Learning] [Perturbation] BioM-JEPA: Joint-Embedding Prediction of Graph-Connected Gene Blocks in Single Cells (arXiv 2026) [preprint]

• [CellWorld] [JEPA] [Foundation Model] [Spatial] [Representation Learning] CellWorld: From Gene-Level Reconstruction to Latent Cell Prediction in Spatial Transcriptomics Foundation Models (arXiv 2026) [preprint] [code]

• [TERRA] [JEPA] [Foundation Model] [Spatial] [Perturbation] [Representation Learning] Multi-Scale Modeling of Human Tissues from Spatial Transcriptomics with TERRA (bioRxiv 2026) [preprint] [code]

• [Context Not Scale] [Virtual Cell] [Review] [Perturbation] Position: Virtual Cells Need Context, Not Just Scale (ICML 2026) [paper] [related implementation]

• [scDifformer] [Virtual Cell] [Foundation Model] [Representation Learning] [Perturbation] scDifformer: diffusion-based post-training for virtual cell modeling across large-scale single-cell data (Nucleic Acids Research 2026) [paper] [code] [project]

• [OCellus] [Virtual Cell] [Foundation Model] [Multimodal] [Spatial] [Perturbation] OCellus: A Language-Model Framework for Single-Cell, Spatial, and Perturbation Biology with Natural-Language Reasoning (bioRxiv 2026) [preprint]

• [CellQ / PACE] [Virtual Cell] [Agent] [Intervention Design] [Perturbation] Virtual-cell verification enables self-auditing AI discovery for immune rejuvenation (bioRxiv 2026) [preprint]

• [VCHarness] [Agent] [Virtual Cell] Harnessing AI to Build Virtual Cells (bioRxiv 2026) [preprint] [code]

• [VCR-Agent] [Agent] [Virtual Cell] [Perturbation] [Gene Regulation] Towards Autonomous Mechanistic Reasoning in Virtual Cells (arXiv 2026) [preprint] [code]

• [SpaCellAgent] [Agent] [Dynamics] [Spatial] [Tool] SpaCellAgent: A Self-Evolving LLM-Based Multi-Agent Framework for Trajectory Analysis (KDD 2026) [paper] [preprint] [code]

• [CellConsensus] [Agent] [Tool] [Dataset] CellConsensus: An agent-curated atlas for automatic cell typing (bioRxiv 2026) [preprint] [code]

• [Literature-Authored Embeddings] [Agent] [Representation Learning] Coding agents author interpretable single-cell embedding models from the literature (bioRxiv 2026) [preprint] [code]

• [LLM4Cell] [Benchmark] [Agent] [Review] LLM4Cell: Taxonomy and Evaluation of LLM and Agentic Models for Single-Cell Biology (ACL 2026) [paper]

• [scBench-Long] [Benchmark] [Agent] [Multimodal] scBench-Long: Verifiable Benchmarking of Long-Horizon Single-Cell Biology (arXiv 2026) [preprint] [code]

• [Score Distributions] [Benchmark] [Perturbation] [Evaluation & Measurement] Score Distributions, Not Cells: Evaluating Single-Cell Perturbations Under Class Overlap (arXiv 2026) [preprint]

• [Projection Basis] [Benchmark] [Perturbation] [Representation Learning] [Evaluation & Measurement] The projection basis determines the information ceiling for perturbation prediction (bioRxiv 2026) [preprint]

• [Harmonised FM Benchmark] [Benchmark] [Foundation Model] [Spatial] [Perturbation] Harmonised benchmarking of foundation models for single-cell and spatial transcriptomics reveals context-dependent generalisation (arXiv 2026) [preprint] [code]

• [scContam] [Benchmark] [Foundation Model] [Evaluation & Measurement] Auditing pretraining contamination in single-cell foundation model benchmarks (arXiv 2026) [preprint] [code]

• [PerturbReason] [Benchmark] [Perturbation] [Agent] [Gene Regulation] PerturbReason: A Knowledge-Grounded Benchmark and Framework for Cell-State–Conditioned Mechanistic Reasoning of Perturbation Effects (NeurIPS 2026) [paper] [preprint] [dataset]

• [DE Classification] [Benchmark] [Perturbation] [Evaluation & Measurement] Beyond Expression Prediction: Benchmarking Differential Expression Classification in Single-Cell Perturbation Models (bioRxiv 2026) [preprint]

• [Principled Evaluation] [Benchmark] [Perturbation] [Evaluation & Measurement] Towards Principled Evaluation of Single-Cell Perturbation Prediction Models (bioRxiv 2026) [preprint] [code]

• [Response Magnitude] [Benchmark] [Perturbation] [Foundation Model] [Evaluation & Measurement] Response Magnitude as a Dominant Signal for Held-Out CRISPRi Perturbation Effect Prediction (arXiv 2026) [preprint]

• [SAFFRON] [Benchmark] [Foundation Model] [Spatial] Evaluating the ability of spatial transcriptomics foundation models to learn multi-scale spatial variation (bioRxiv 2026) [preprint] [code]

• [Confound Diagnostics] [Benchmark] [Foundation Model] [Perturbation] [Tool] [Evaluation & Measurement] A confound-diagnostic toolkit for in silico perturbation with single-cell foundation models (bioRxiv 2026) [preprint] [code]

• [Reliable Perturbations] [Benchmark] [Perturbation] [Evaluation & Measurement] Reliable single-cell perturbations explain and improve model performance (bioRxiv 2026) [preprint]

• [Cell Line Bottleneck] [Benchmark] [Perturbation] [Evaluation & Measurement] Cancer Cell Line Heterogeneity Imposes a Primary Bottleneck for Virtual Perturbation Screening at Scale (bioRxiv 2026) [preprint] [code]

• [ST Agent Benchmark] [Benchmark] [Agent] [Spatial] Mind the alignment gap: a spatial transcriptomics benchmark for scientific coding agents (bioRxiv 2026) [preprint]

• [CRISPRko vs CRISPRi] [Benchmark] [Dataset] [Perturbation] Direct comparison of CRISPR knockout and interference with Perturb-seq (bioRxiv 2026) [preprint] [code]

• [JUMP-lite] [Benchmark] [Morphology] [Dataset] [Tool] JUMP-lite: Compact, reproducible benchmarking of cell representations (arXiv 2026) [preprint] [code]

• [scVision] [Foundation Model] [Representation Learning] A vision foundation model for single-cell biology via spatial gene cartography (arXiv 2026) [preprint] [project]

• [SATScG] [Foundation Model] [Representation Learning] Scaling an Autoregressive Transformer for Single-Cell Generation (arXiv 2026) [preprint] [code]

• [Gene Intelligence] [Foundation Model] [Representation Learning] [Benchmark] Raw-count embeddings improve single-cell foundation models (bioRxiv 2026) [preprint] [code]

• [Task-Adapted FM] [Foundation Model] [Perturbation] [Representation Learning] Task-adapted biological foundation models uncover perturbation-centric representations (bioRxiv 2026) [preprint] [code]

• [Species-Native Tokens] [Foundation Model] [Representation Learning] Single-cell foundation modeling with species-native protein tokens links regenerative competence across frog and mouse (bioRxiv 2026) [preprint] [code]

• [Complementary Views] [Foundation Model] [Representation Learning] [Gene Regulation] Beyond Gene Reconstruction: Learning Cell Representations through Complementary Transcriptomic Views (arXiv 2026) [preprint]

• [Tabula] [Foundation Model] [Gene Regulation] [Intervention Design] Predictive single cell foundation model for gene regulation and aging with privacy-preserving tabular learning (arXiv 2026) [preprint] [code]

• [AdaGeneBudget] [Foundation Model] [Tool] [Representation Learning] AdaGeneBudget: Cell-Adaptive Gene-Token Allocation for Efficient Single-Cell Foundation Models (bioRxiv 2026) [preprint]

• [CellTosg2Sequence] [Foundation Model] [Multimodal] [Representation Learning] CellTosg2Sequence: A Unified Text-Omics-Signaling-Graph Large Language Model for Single-Cell Analysis (bioRxiv 2026) [preprint]

• [Stable-Shift] [Perturbation] Stable-Shift: Predicting Transcriptional Responses of Unseen Gene Perturbations Using Graph Neural Networks with Biological Priors (ACM BCB 2026) [paper] [preprint] [code]

• [PertOmni] [Perturbation] [Multimodal] [Morphology] [Representation Learning] Learning Perturbation Effects Through Contrastive Alignment of Multimodal Biological Embeddings (bioRxiv 2026) [preprint]

• [scCycleMol] [Perturbation] Modeling Cell-Cycle-Aware Single-Cell Drug Perturbation Responses (arXiv 2026) [preprint]

• [U-Pert] [Perturbation] [Dynamics] [Intervention Design] Unbalanced Perturbation Dynamics For Cell Fate Design (bioRxiv 2026) [preprint] [code]

• [GeneSpeak-FP] [Perturbation] [Intervention Design] GeneSpeak-FP: Target and Compound Retrieval from Observed Cell-Level Perturbation Signatures (arXiv 2026) [preprint]

• [PerturbPFN] [Perturbation] [Foundation Model] [Gene Regulation] PerturbPFN: Probing the Limits of Synthetic Priors in Drug Perturbation Modelling (ICML 2026) [paper]

• [Response Decomposition] [Perturbation] [Benchmark] Perturbation response decomposition enables biologically aligned generalization to unseen perturbations and cellular contexts (bioRxiv 2026) [preprint] [code]

• [Cytokine Atlas] [Perturbation] [Gene Regulation] [Dataset] A human cytokine response atlas to reconstruct underlying gene regulatory networks (bioRxiv 2026) [preprint] [code] [project] [dataset]

• [PerturbMap] [Perturbation] PerturbMap: Cross-Context Transfer of Single-Cell Perturbation Responses (arXiv 2026) [preprint]

• [LGR] [Perturbation] [Agent] LLM-Guided Retrieval for Prediction of Molecular Perturbation Responses (ICLR 2026) [paper]

• [MEGA-ODE] [Perturbation] [Dynamics] [Gene Regulation] [Intervention Design] MEGA-ODE: Learning Biologically Structured and Navigable Continuous Perturbation Dynamics from Sparse Omics (bioRxiv 2026) [preprint] [code]

• [TranScouter] [Perturbation] [Benchmark] A structured study of cross-condition prediction of transcriptional responses to gene perturbations (bioRxiv 2026) [preprint] [code]

• [COMPASS] [Perturbation] COMPASS: Component-Wise Inference of Shared and Gene-Specific Perturbation Response (bioRxiv 2026) [preprint] [code]

• [SLIM] [Perturbation] SLIM: A small linear model with STRING embeddings for single-cell genetic perturbation prediction (bioRxiv 2026) [preprint] [code]

• [GeneGeoFlow] [Perturbation] Control-Anchored Residual Flow Matching Conditioned on Gene Geometry for Virtual Cell Perturbation Modeling (arXiv 2026) [preprint]

• [PerturbLDM] [Perturbation] PerturbLDM: conditional latent diffusion for modelling single-cell perturbation responses (bioRxiv 2026) [preprint] [code]

• [Coladan] [Spatial] [Morphology] [Multimodal] [Perturbation] [Dataset] Trimodal, uncertainty-guided whole-slide framework for genome-scale spatial expression and image-only virtual perturbation in cancer cohorts (Genome Medicine 2026) [paper] [code]

• [Spaceland] [Spatial] [Morphology] [Multimodal] Spaceland: Histology-Guided Reconstruction of High-Resolution Whole-Organ 3D Molecular Atlases from Sparse Spatial Transcriptomics (bioRxiv 2026) [preprint]

• [TissueFormer (Pathology)] [Spatial] [Morphology] [Multimodal] [Foundation Model] Multi-Modal Foundation Model with Whole-Slide Attention Enables Transferrable Digital Pathology at Single-Cell Resolution (bioRxiv 2026) [preprint] [code]

• [VOICE] [Spatial] [Morphology] [Multimodal] [Foundation Model] VOICE: A Vision-Omics Foundation Model Integrating Direct and Retrieval-Based Prediction of In-situ Single-Cell Gene Expression (arXiv 2026) [preprint] [code]

• [VISTA] [Spatial] [Morphology] [Multimodal] Virtual spatial transcriptomics from histopathology enables prognostic and therapeutic response prediction in cancer (bioRxiv 2026) [preprint] [code]

• [SQUINT] [Spatial] [Representation Learning] Learning Discrete Cell and Niche Codes from Spatial Transcriptomics Using Dual Residual Vector Quantization (bioRxiv 2026) [preprint] [code (project)]

• [MAE-3D] [Morphology] [Multimodal] [Representation Learning] [Benchmark] 3D Masked Autoencoders are Robust Learners of Volumetric and Multimodal Cellular Representations for Microscopy (arXiv 2026) [preprint] [code]

• [Cell Painting RAG Audit] [Morphology] [Agent] [Benchmark] Auditing Retrieval-Augmented LLM Hypotheses for Longitudinal Cell Painting Morphology (ACM BCB 2026) [paper] [preprint]

• [Spatium] [Protein] [Spatial] [Foundation Model] [Representation Learning] Spatium: A Protein Language Foundation Model for Spatial Proteomics (bioRxiv 2026) [preprint] [code]

• [PerturbMatch] [Perturbation] [Tool] Joint analysis of multiply perturbed cells improves statistical power and cost efficiency in Perturb-seq (bioRxiv 2026) [preprint] [code]

• [scRepresenter] [Representation Learning] [Tool] [Benchmark] [Foundation Model] scRepresenter: a workflow for computing, integrating and benchmarking cellular representations in single-cell transcriptomics (bioRxiv 2026) [preprint] [code]

• [CELLens] [Virtual Cell] [Gene Regulation] [Tool] Human-Guided Causal Knowledge Injection for Virtual Cells (arXiv 2026) [preprint] [code]

• [Tabular FM Perturbation] [Benchmark] [Perturbation] [Foundation Model] Tabular Foundation Models Are Competitive Cellular Perturbation Predictors Across Biological Scales (bioRxiv 2026) [preprint] [code]

• [CellFM-Datasets] [Foundation Model] [Spatial] [Tool] Cellfm-datasets: A Unified Data Infrastructure for Single-Cell and Spatial Transcriptomics Foundation Model Pretraining (bioRxiv 2026) [preprint]

• [OCOO-T] [Virtual Cell] [Perturbation] OCOO-T : A Simple and Scalable Virtual Cell Model for Transcriptional Perturbation Response Prediction (arXiv 2026) [preprint]

• [Glitch Genes] [Benchmark] [Foundation Model] [Representation Learning] Do Geometric Outliers Identify Important Genes in Single-Cell Foundation Models? (bioRxiv 2026) [preprint]

• [VCBench] [Virtual Cell] [Benchmark] [Foundation Model] VCBench: A Multi-Dimensional Benchmark for Single-Cell Foundation Models (bioRxiv 2026) [preprint] [code]

• [Design Space] [Perturbation] [Benchmark] [Representation Learning] Elucidating the Design Space of Generative Models for Single-Cell Perturbation Prediction (bioRxiv 2026) [preprint] [code]

• [PertDiffBench] [Perturbation] [Benchmark] PertDiffBench: Benchmarking Diffusion Models for Single-Cell Perturbation Response Prediction (bioRxiv 2026) [preprint] [code (project)]

• [Zero-Shot Benchmark] [Benchmark] [Foundation Model] [Representation Learning] Context-dependent utility and robustness of pretrained single-cell foundation model representations across analytical tasks (bioRxiv 2026) [preprint] [code]

• [DeepSpot-M] [Spatial] [Morphology] [Multimodal] [Foundation Model] DeepSpot-M: a multimodal foundation model for transcriptome-wide virtual spatial transcriptomics from histology (medRxiv 2026) [preprint] [code]

• [V3Cell] [Virtual Cell] [Morphology] [Perturbation] [Dynamics] V3Cell: A Vision-Guided Virtual 3D Cell Framework for Phenotypic Modeling and Perturbation Prediction (bioRxiv 2026) [preprint] [code]

• [Cross-Context DrugPert] [Perturbation] [Multimodal] Enhancing Cross-Context Generalization in Drug Perturbation Prediction with a Multimodal Conditional Diffusion Framework (Bioinformatics 2026) [paper] [code]

• [SciCore-Omics] [Foundation Model] [Spatial] [Morphology] [Multimodal] SciCore-Omics: a tri-modal foundation model unifying histology, spatial transcriptomics and language for spatial biology (bioRxiv 2026) [preprint] [code]

• [HoloCell] [Virtual Cell] [Foundation Model] [Multimodal] [Protein] HoloCell: A Generative Foundation Model for Holistic Cellular Modeling (bioRxiv 2026) [preprint] [project (code pending)]

• [FM Roadmap] [Foundation Model] [Spatial] [Review] A User’s Roadmap to Foundation Models on Single-Cell and Spatial-Omics – Cell Type and Lineage applications (National Science Review 2026) [paper]

• [PerturbCellRL] [Perturbation] PerturbCellRL: Verifier-Guided Reinforcement Learning for Single-Cell Perturbation Prediction (arXiv 2026) [preprint]

• [KG-Reasoning LLM] [Perturbation] [Agent] Knowledge Graphs and Reasoning LLMs for Finding Simple Yet Effective Transcriptomic Perturbation Predictors (arXiv 2026) [preprint]

• [Cross-Modal Transfer] [Foundation Model] [Multimodal] [Spatial] Single-Cell Cross-Modal Transfer by Adversarial Fine-Tuning of Foundation Models (arXiv 2026) [preprint]

• [BRIDGE] [Foundation Model] [Morphology] [Spatial] [Multimodal] BRIDGE: A Multi-organ Histo-ST Foundation Model Enables Virtual Spatial Transcriptomics for Enhanced Few-shot Cancer Diagnosis (bioRxiv 2026) [preprint] [code]

• [Morphodynamics-Expr] [Morphology] [Dynamics] [Multimodal] Single-cell morphodynamical trajectories enable prediction of gene expression accompanying cell state change (Cell Systems 2026) [paper] [code] [implementation package]

• [Morph-Transcriptomic GenModel] [Perturbation] [Morphology] [Multimodal] A generative framework for predicting cellular morphological and transcriptomic perturbation responses (Cell Reports Methods 2026) [paper] [code]

• [DoFormer] [Perturbation] [Gene Regulation] [Multimodal] DoFormer: Causal Transformer for Gene Perturbation (bioRxiv 2026) [preprint]

• [RegFormer] [Foundation Model] [Gene Regulation] [Representation Learning] RegFormer: a single-cell foundation model powered by gene regulatory hierarchies (Nature Communications 2026) [paper] [code]

• [Spurious Correlation] [Benchmark] [Perturbation] [Evaluation & Measurement] Spurious correlation inflates performance in single-cell perturbation prediction (bioRxiv 2026) [preprint] [code]

• [scArchon] [Benchmark] [Perturbation] [Tool] scArchon: a scalable benchmarking framework for assessing single-cell perturbation models (Genome Biology 2026) [paper] [code]

• [Chemical Pert DL] [Benchmark] [Perturbation] [Evaluation & Measurement] Deep learning models for chemical perturbation prediction do not yet utilise drug molecular features (bioRxiv 2026) [preprint]

• [Cycle-Consistent GenModel] [Spatial] [Multimodal] [Protein] [Representation Learning] Cycle-consistent deep generative modeling unifies cellular states across unpaired spatial and single-cell modalities (bioRxiv 2026) [preprint] [code]

• [StateXDiff] [Perturbation] [Multimodal] StateXDiff: Cell State-Contextualized Multimodal Diffusion for Single-Cell Perturbation Prediction (arXiv 2026) [preprint]

• [DeSCOPE] [Perturbation] [Multimodal] Decoding Single-Cell Omics of Perturbation Responses Using DeSCOPE (bioRxiv 2026) [preprint] [code]

• [Dataset Size & Diversity] [Foundation Model] [Benchmark] Evaluating the role of pretraining dataset size and diversity on single-cell foundation model performance (Nature Methods 2026) [paper] [code]

• [Lingshu-Cell] [Virtual Cell] [World Model] [Foundation Model] [Perturbation] Lingshu-Cell: A generative cellular world model for transcriptome modeling toward virtual cells (arXiv 2026) [preprint] [code] [homepage]

• [SCALE] [Perturbation] SCALE: Scalable Conditional Atlas-Level Endpoint transport for virtual cell perturbation prediction (arXiv 2026) [preprint]

• [Conditional Monge Gap] [Perturbation] Conditional Monge Gap enables generalizable single-cell perturbation modelling (Nature Machine Intelligence 2026) [paper] [code]

• [ProtiCelli] [Virtual Cell] [Protein] [Morphology] Generative machine learning unlocks the first proteome-wide image of human cells (bioRxiv 2026) [preprint] [code]

• [AetherCell] [Virtual Cell] [Foundation Model] [Perturbation] [Intervention Design] AetherCell: A generative engine for virtual cell perturbation and in vivo drug discovery (bioRxiv 2026) [preprint] [code]

• [AlphaCell] [Virtual Cell] [World Model] [Perturbation] [Dynamics] Towards building a World Model to simulate perturbation-induced cellular dynamics by AlphaCell (bioRxiv 2026) [preprint]

• [VCWorld] [Virtual Cell] [World Model] [Foundation Model] [Perturbation] VCWorld: A Biological World Model for Virtual Cell Simulation (ICLR 2026) [paper] [code] [ask deepwiki]

• [Spatial Perturb-seq] [Perturbation] [Spatial] [Dataset] Spatial perturb-seq: single-cell functional genomics within intact tissue architecture (Nature Communications 2026) [paper] [code]

• [Celcomen] [Perturbation] [Spatial] [Gene Regulation] Celcomen: spatial causal disentanglement for single-cell and tissue perturbation modeling (Nature Communications 2026) [paper] [code]

• [stVCR] [Spatial] [Dynamics] stVCR: spatiotemporal dynamics of single cells (Nature Methods 2026) [paper] [code]

• [CONCORD] [Representation Learning] Revealing a coherent cell-state landscape across single-cell datasets with CONCORD (Nature Biotechnology 2026) [paper] [code]

• [AI Scientist] [Agent] [Related] Towards end-to-end automation of AI research (Nature 2026) [paper] [code] [template-free code]

• [Conformation Description Language] [Protein] [Multimodal] [Related] Bridging three-dimensional molecular structures and artificial intelligence with a conformation description language (Nature Machine Intelligence 2026) [paper] [code]

• [CRISPRi Map] [Perturbation] [Gene Regulation] [Dataset] A genome-scale single-cell CRISPRi map of trans gene regulation across human pluripotent stem cell lines (Cell Genomics 2026) [paper] [code] [publisher]

• [TxPert] [Perturbation] TxPert: using multiple knowledge graphs for prediction of transcriptomic perturbation effects (Nature Biotechnology 2026) [paper] [code]

• [Therapeutic Design] [Perturbation] [Intervention Design] Deep-learning-based de novo discovery and design of therapeutics that reverse disease-associated transcriptional phenotypes (Cell 2026) [paper] [code]

• [X-Pert] [Perturbation] [Multimodal] Unified Multimodal Learning Enables Generalized Cellular Response Prediction to Diverse Perturbations (bioRxiv 2026) [preprint] [code] [ask deepwiki]

• [MVCBench] [Benchmark] [Perturbation] [Multimodal] [Morphology] MVCBench: A Multimodal Benchmark for Drug-induced Virtual Cell Phenotypes (bioRxiv 2026) [preprint] [code] [ask deepwiki]

• [HarmonyCell] [Agent] [Perturbation] [Tool] HarmonyCell: Automating Single-Cell Perturbation Modeling under Semantic and Distribution Shifts (bioRxiv 2026) [preprint]

• [scDFM] [Perturbation] scDFM: Distributional Flow Matching Model for Robust Single-Cell Perturbation Prediction (ICLR 2026) [paper] [code]

• [Doloris] [Perturbation] Doloris: Dual Conditional Diffusion Implicit Bridges with Sparsity Masking Strategy for Unpaired Single-Cell Perturbation Estimation (ICLR 2026) [paper] [code]

• [Departures] [Perturbation] Departures: Distributional Transport for Single-Cell Perturbation Prediction with Neural Schrödinger Bridges (AAAI 2026) [paper] [preprint] [code]

• [PETRI] [Foundation Model] [Multimodal] [Representation Learning] PETRI: Learning Unified Cell Embeddings from Unpaired Modalities via Early-Fusion Joint Reconstruction (ICLR 2026) [paper]

• [STRAND] [Perturbation] [Gene Regulation] STRAND: Sequence-Conditioned Transport for Single-Cell Perturbations (NeurIPS 2026) [paper] [preprint]

• [PerturbDiff] [Perturbation] PerturbDiff: Functional Diffusion for Single-Cell Perturbation Modeling (ICML 2026) [paper] [preprint] [code] [project]

• [scBIG] [Perturbation] [Representation Learning] [Gene Regulation] Beyond Independent Genes: Learning Module-Inductive Representations for Single-Cell Gene Perturbation Prediction (ICML 2026) [paper] [preprint] [code]

• [CellxPert] [Foundation Model] [Perturbation] [Multimodal] [Spatial] [Protein] CellxPert: Inference-Time MCMC Steering of a Multi-Omics Single-Cell Foundation Model for In-Silico Perturbation (ICLR 2026) [paper]

• [Perturbation Representation] [Perturbation] [Representation Learning] [Benchmark] [Evaluation & Measurement] What Makes a Representation Good for Single-Cell Perturbation Prediction? (ICML 2026) [paper] [preprint] [code]

• [CisTransCell] [Perturbation] [Gene Regulation] [Multimodal] CisTransCell: Single-Cell Perturbation Prediction via Gene Function, Regulatory Control, and Cellular Context (arXiv 2026) [preprint]

• [Latent Causal Processes] [Perturbation] [Dynamics] [Gene Regulation] Learning Latent Dynamical Causal Processes for Single-Cell Perturbation Prediction (KDD 2026) [paper] [preprint] [code]

• [msInfer] [Protein] [Multimodal] [Tool] Large-scale proteome inference from unpaired single-cell transcriptomic and proteomic data by msInfer (Research Square 2026) [preprint] [code]

• [Stack] [Foundation Model] [Representation Learning] [Perturbation] Stack: In-Context Learning of Single-Cell Biology (bioRxiv 2026) [preprint] [code] [ask deepwiki]

• [BioWorldModel] [World Model] [Dynamics] [Related] BioWorldModel: a single architecture predicts phenotype from genotype across four kingdoms of life (bioRxiv 2026) [preprint]

• [Gene Importance] [Foundation Model] [Gene Regulation] [Tool] Scoring gene importance by interpreting single-cell foundation models (Nature Biotechnology 2026) [paper] [code]

• [Hi-C FM] [Foundation Model] [Gene Regulation] [Multimodal] A generalizable Hi-C foundation model for chromatin architecture, single-cell and multiomics analysis across species (Nature Methods 2026) [paper] [code]

• [Virtual Spatial Tumor] [Morphology] [Spatial] [Protein] [Multimodal] Cellular architecture and neighborhood-informed virtual spatial tumor profiling from histopathology (Cell 2026) [paper] [code]

• [SynCell] [Virtual Cell] [Review] [Related] A framework for building a synthetic cell from the SynCell Asia Initiative (Nature Biotechnology 2026) [paper]

• [3D Genome FM] [Foundation Model] [Gene Regulation] [Review] A foundation model to help understand the regulatory implications of 3D genome organization (Nature Methods 2026) [paper]

• [CytoSignal] [Spatial] [Dynamics] [Gene Regulation] [Tool] CytoSignal detects locations and dynamics of ligand–receptor signaling at cellular resolution from spatial transcriptomic data (Nature Genetics 2026) [paper] [code]

• [graphene-seq] [Spatial] [Multimodal] [Dynamics] [Dataset] In situ graphene-seq: spatial transcriptomics and chronic electrophysiological characterization of tissue microenvironments (Nature Communications 2026) [paper] [code]

• [Deep Molecular Profiling] [Spatial] [Multimodal] [Review] Deep molecular profiling in three dimensions (Nature Methods 2026) [paper]

• [SpaMosaic] [Spatial] [Multimodal] [Protein] [Representation Learning] Mosaic integration of spatial multi-omics with SpaMosaic (Nature Genetics 2026) [paper] [code]

• [veloAgent] [Dynamics] [Spatial] [Intervention Design] Dissecting and steering cell dynamics using spatially-informed RNA velocity with veloAgent (Molecular Systems Biology 2026) [paper] [code]

• [Morphodynamics] [Morphology] [Dynamics] Single-cell morphodynamics predict cell fate decisions during mucociliary epithelial differentiation (Molecular Systems Biology 2026) [paper] [code] [image-processing code]

• [Single Cell Notebooks] [Tool] [Spatial] The Single Cell Notebooks for inclusive and accessible training in single-cell and spatial omics (Nature Genetics 2026) [paper] [code]

• [CAPTAIN] [Foundation Model] [Multimodal] [Protein] [Representation Learning] CAPTAIN: a multimodal foundation model pretrained on co-assayed single-cell RNA and protein (Nature Communications 2026) [paper] [preprint] [code]

• [scpFormer] [Foundation Model] [Protein] [Representation Learning] scpFormer: A Foundation Model for Unified Representation and Integration of the Single-Cell Proteomics (arXiv 2026) [preprint] [code]

• [MultiPert] [Perturbation] [Multimodal] [Protein] MultiPert: An adversarial alignment and dual attention framework for single-cell multi-omics perturbation prediction (PLOS Computational Biology 2026) [paper] [code]

• [UniCure] [Foundation Model] [Perturbation] [Multimodal] [Intervention Design] UniCure: A multi-modal model for predicting personalized cancer therapy response (Cancer Cell 2026) [paper] [preprint] [code] [ask deepwiki]

🗓️ 2025 — 123 papers

• [CASSIA] [Agent] [Tool] CASSIA: a multi-agent large language model for automated and interpretable cell annotation (Nature Communications 2025) [paper] [preprint] [code]

• [CP-POSH] [Morphology] [Perturbation] [Representation Learning] A pooled Cell Painting CRISPR screening platform enables de novo inference of gene function by self-supervised deep learning (Nature Communications 2025) [paper] [preprint] [code] [models] [data]

• [inspre] [Gene Regulation] [Perturbation] Large-scale causal discovery using interventional data sheds light on gene network structure in k562 cells (Nature Communications 2025) [paper] [preprint] [code] [data]

• [MorphoGenie] [Morphology] [Representation Learning] Generalizable morphological profiling of cells by interpretable unsupervised learning (Nature Communications 2025) [paper] [preprint] [code]

• [scFM Biology Insights] [Foundation Model] [Benchmark] [Evaluation & Measurement] Biology-driven insights into the power of single-cell foundation models (Genome Biology 2025) [paper] [code]

• [scKGBERT] [Foundation Model] [Representation Learning] scKGBERT: a knowledge-enhanced foundation model for single-cell transcriptomics (Genome Biology 2025) [paper] [preprint] [models] [dataset]

• [SIDISH] [Perturbation] [Intervention Design] [Spatial] SIDISH integrates single-cell and bulk transcriptomics to identify high-risk cells and guide precision therapeutics through in silico perturbation (Nature Communications 2025) [paper] [preprint] [code]

• [STORIES] [Spatial] [Dynamics] STORIES: learning cell fate landscapes from spatial transcriptomics using optimal transport (Nature Methods 2025) [paper] [preprint] [code] [reproducibility]

• [CONCERT] [Spatial] [Perturbation] [Dynamics] Generative modeling of spatially resolved perturbation responses with CONCERT (bioRxiv 2025) [preprint] [code] [project]

• [CIPHER] [Perturbation] [Gene Regulation] [Intervention Design] Co-fluctuations of genes in single cells predict transcriptome-wide outcomes to perturbations (bioRxiv 2025) [preprint] [code] [project] [data] [data]

• [NetDes] [Gene Regulation] [Dynamics] Building dynamical models of multi-step state transitions from single cell gene expression trajectories (bioRxiv 2025) [preprint] [code] [reproducibility]

• [scConcept] [Foundation Model] [Representation Learning] [Spatial] scConcept: Contrastive pretraining for technology-agnostic single-cell representations beyond reconstruction (bioRxiv 2025) [preprint] [code] [documentation]

• [scFM Cancer Outcomes] [Foundation Model] [Benchmark] [Evaluation & Measurement] [Representation Learning] [Agent] Empirical Evaluation of Single-Cell Foundation Models for Predicting Cancer Outcomes (bioRxiv 2025) [preprint] [code]

• [scFME] [Benchmark] [Foundation Model] [Perturbation] [Evaluation & Measurement] Single Cell Foundation Models Evaluation (scFME) for In-Silico Perturbation (bioRxiv 2025) [preprint]

• [PEREGGRN] [Benchmark] [Perturbation] [Gene Regulation] [Evaluation & Measurement] A comparison of computational methods for expression forecasting (Genome Biology 2025) [paper] [preprint] [code]

• [PertEval-scFM] [Benchmark] [Foundation Model] [Perturbation] [Evaluation & Measurement] PertEval-scFM: Benchmarking Single-Cell Foundation Models for Perturbation Effect Prediction (ICML 2025) [paper] [preprint] [code]

• [BioDiscoveryAgent] [Agent] [Intervention Design] [Perturbation] BioDiscoveryAgent: An AI Agent for Designing Genetic Perturbation Experiments (ICLR 2025) [paper] [preprint] [code]

• [PerturbQA] [Benchmark] [Perturbation] [Agent] Contextualizing biological perturbation experiments through language (ICLR 2025) [paper] [preprint] [code]

• [dbDiffusion] [Perturbation] Predicting the unseen: A diffusion-based debiasing framework for transcriptional response prediction at single-cell resolution (PNAS 2025) [paper] [code]

• [scCausalVI] [Perturbation] [Representation Learning] scCausalVI disentangles single-cell perturbation responses with causality-aware generative model (Cell Systems 2025) [paper] [preprint] [code]

• [CellCap] [Perturbation] [Representation Learning] Explainable modeling of single-cell perturbation data using attention and sparse dictionary learning (Cell Systems 2025) [paper] [preprint] [code (project)]

• [Meta Flow Matching] [Perturbation] [Dynamics] Meta Flow Matching: Integrating Vector Fields on the Wasserstein Manifold (ICLR 2025) [paper] [preprint] [code]

• [scFM Zero-Shot Evaluation] [Benchmark] [Evaluation & Measurement] [Foundation Model] [Representation Learning] Zero-shot evaluation reveals limitations of single-cell foundation models (Genome Biology 2025) [paper] [preprint] [code]

• [Novae] [Foundation Model] [Spatial] [Representation Learning] Novae: a graph-based foundation model for spatial transcriptomics data (Nature Methods 2025) [paper] [preprint] [code]

• [scooby] [Foundation Model] [Gene Regulation] [Multimodal] scooby: modeling multimodal genomic profiles from DNA sequence at single-cell resolution (Nature Methods 2025) [paper] [preprint] [code (project)]

• [ViTally Consistent] [Foundation Model] [Morphology] [Representation Learning] ViTally Consistent: Scaling Biological Representation Learning for Cell Microscopy (ICML 2025) [paper] [preprint] [code (project)]

• [Cell-DINO] [Foundation Model] [Morphology] [Representation Learning] Cell-DINO: Self-supervised image-based embeddings for cell fluorescent microscopy (PLOS Computational Biology 2025) [paper] [code] [Cell-DINO docs]

• [ChromFound] [Foundation Model] [Gene Regulation] [Representation Learning] ChromFound: Towards A Universal Foundation Model for Single-Cell Chromatin Accessibility Data (NeurIPS 2025) [paper] [preprint] [code]

• [GHIST] [Spatial] [Morphology] [Multimodal] Spatial gene expression at single-cell resolution from histology using deep learning with GHIST (Nature Methods 2025) [paper] [preprint] [code]

• [GigaTIME] [Spatial] [Protein] [Morphology] [Multimodal] Multimodal AI generates virtual population for tumor microenvironment modeling (Cell 2025) [paper] [code]

• [PUPS] [Protein] [Morphology] [Multimodal] Prediction of protein subcellular localization in single cells (Nature Methods 2025) [paper] [preprint] [code (project)]

• [KRONOS] [Foundation Model] [Spatial] [Protein] [Representation Learning] A Foundation Model for Spatial Proteomics (arXiv 2025) [preprint] [code]

• [PerTurboAgent] [Agent] [Perturbation] [Intervention Design] PerTurboAgent: An LLM-based Agent for Designing Iterative Perturb-Seq Experiments (MLCB 2025) [paper] [preprint]

• [CellVerse] [Agent] [Benchmark] [Evaluation & Measurement] CellVerse: Do Large Language Models Really Understand Cell Biology? (NeurIPS 2025) [paper] [preprint]

• [Virtual Lab] [Agent] [Protein] [Related] The Virtual Lab of AI agents designs new SARS-CoV-2 nanobodies (Nature 2025) [paper] [preprint] [code]

• [LLMs Meet Virtual Cell] [Review] [Agent] [Virtual Cell] Large Language Models Meet Virtual Cell: A Survey (arXiv 2025) [preprint]

• [Human Cytokine Dictionary] [Dataset] [Perturbation] A single-cell cytokine dictionary of human peripheral blood (bioRxiv 2025) [preprint] [Parse 10M PBMC dataset]

• [Open Problems] [Benchmark] [Evaluation & Measurement] [Tool] Defining and benchmarking open problems in single-cell analysis (Nature Biotechnology 2025) [paper] [code (project)] [platform]

• [CRISP] [Foundation Model] [Perturbation] Predicting drug responses of unseen cell types through transfer learning with foundation models (Nature Computational Science 2025) [paper] [code]

• [PDGrapher] [Perturbation] [Intervention Design] [Gene Regulation] [Representation Learning] Combinatorial prediction of therapeutic perturbations using causally inspired neural networks (Nature Biomedical Engineering 2025) [paper] [preprint] [code] [project]

• [PAIRING] [Perturbation] [Intervention Design] [Representation Learning] Identifying an optimal perturbation to induce a desired cell state by generative deep learning (Cell Systems 2025) [paper] [code]

• [ARC] [Intervention Design] [Gene Regulation] [Dynamics] [Related] Reverse control of biological networks to restore phenotype landscapes (Science Advances 2025) [paper] [code] [software archive]

• [PerturbNet] [Perturbation] [Intervention Design] [Representation Learning] PerturbNet predicts single-cell responses to unseen chemical and genetic perturbations (Molecular Systems Biology 2025) [paper] [preprint] [code]

• [GeneJEPA] [JEPA] [World Model] [Foundation Model] [Representation Learning] [Perturbation] GeneJEPA: A Predictive World Model of the Transcriptome (bioRxiv 2025) [preprint] [code]

• [STELLA] [Agent] [Multimodal] [Related] STELLA: Self-Evolving LLM Agent for Biomedical Research (bioRxiv 2025) [preprint] [code]

• [scPRINT-2] [Foundation Model] [Representation Learning] [Perturbation] [Gene Regulation] [Benchmark] scPRINT-2: Towards the Next Generation of Cell Foundation Models and Benchmarks (bioRxiv 2025) [preprint] [code]

• [Pertpy] [Perturbation] [Tool] Pertpy: an End-to-end Framework for Perturbation Analysis (Nature Methods 2025) [paper] [code] [ask deepwiki]

• [scPerturBench (Nat Methods)] [Benchmark] [Perturbation] [Foundation Model] Benchmarking Algorithms for Generalizable Single-Cell Perturbation Response Prediction (Nature Methods 2025) [paper] [code] [ask deepwiki]

• [DeepSpot2Cell] [Spatial] [Morphology] [Multimodal] DeepSpot2Cell: Predicting Virtual Single-Cell Spatial Transcriptomics from H&E images using Spot-Level Supervision (NeurIPS 2025) [paper] [code]

• [Scouter] [Perturbation] Scouter predicts transcriptional responses to genetic perturbations with large language model embeddings (Nature Computational Science 2025) [paper] [code] [reproduce]

• [GPerturb] [Perturbation] GPerturb: Gaussian process modelling of single-cell perturbation data (Nature Communications 2025) [paper] [code]

• [Squidiff] [Perturbation] [Dynamics] Squidiff: Predicting Cellular Development and Responses to Perturbations using a Diffusion Model (Nature Methods 2025) [paper] [code] [ask deepwiki]

• [Nicheformer] [Foundation Model] [Spatial] [Representation Learning] Nicheformer: A Foundation Model for Single-Cell and Spatial Omics (Nature Methods 2025) [paper] [code] [中文解读] [ask deepwiki]

• [NicheCompass] [Spatial] [Gene Regulation] [Representation Learning] Quantitative characterization of cell niches in spatially resolved omics data (Nature Genetics 2025) [paper] [code]

• [STAMP] [Dataset] [Multimodal] [Protein] [Morphology] STAMP: Single-cell transcriptomics analysis and multimodal profiling through imaging (Cell 2025) [paper] [code]

• [Perturb-FISH] [Perturbation] [Spatial] [Dataset] Simultaneous CRISPR screening and spatial transcriptomics reveal intracellular, intercellular, and functional transcriptional circuits (Cell 2025) [paper] [code]

• [ADLF] [Perturbation] [Intervention Design] Active Learning Framework Leveraging Transcriptomics Identifies Modulators of Disease Phenotypes (Science 2025) [paper] [code] [software archive]

• [Tahoe-x1] [Foundation Model] [Perturbation] [Representation Learning] Tahoe-x1: Scaling Perturbation-Trained Single-Cell Foundation Models to 3 Billion Parameters (bioRxiv 2025) [preprint] [code] [ask deepwiki] [hugging face files]

• [LPM] [Foundation Model] [Perturbation] [Multimodal] In Silico Biological Discovery with Large Perturbation Models (Nature Computational Science 2025) [paper] [code] [中文解读] [ask deepwiki]

• [CellNavi] [Perturbation] [Intervention Design] [Gene Regulation] CellNavi Predicts Genes Directing Cellular Transitions by Learning a Gene Graph-Enhanced Cell State Manifold (Nature Cell Biology 2025) [paper] [code] [中文解读] [ask deepwiki]

• [EpiAgent] [Foundation Model] [Representation Learning] [Gene Regulation] EpiAgent: Foundation Model for Single-Cell Epigenomics (Nature Methods 2025) [paper] [code] [中文解读] [ask deepwiki]

• [CellWhisperer] [Foundation Model] [Multimodal] [Representation Learning] [Tool] Multimodal learning enables chat-based exploration of single-cell data (Nature Biotechnology 2025) [paper] [code]

• [CRISPR-GPT] [Agent] [Perturbation] [Intervention Design] [Tool] CRISPR-GPT for Agentic Automation of Gene-Editing Experiments (Nature Biomedical Engineering 2025) [paper] [code] [中文解读] [ask deepwiki]

• [Systema] [Benchmark] [Perturbation] [Evaluation & Measurement] Systema: A Framework for Evaluating Genetic Perturbation Response Prediction Beyond Systematic Variation (Nature Biotechnology 2025) [paper] [code] [中文解读] [ask deepwiki]

• [IMPA] [Perturbation] [Morphology] Predicting cell morphological responses to perturbations using generative modeling (Nature Communications 2025) [paper] [code]

• [PhenoProfiler] [Morphology] [Representation Learning] PhenoProfiler: advancing phenotypic learning for image-based drug discovery (Nature Communications 2025) [paper] [code] [webserver] [ask deepwiki]

• [PERISCOPE] [Morphology] [Perturbation] [Dataset] A genome-wide atlas of human cell morphology (Nature Methods 2025) [paper] [code] [dataset]

• [Morph Map] [Morphology] [Perturbation] [Dataset] Morphological map of under- and overexpression of genes in human cells (Nature Methods 2025) [paper] [code] [dataset]

• [MorphDiff] [Perturbation] [Morphology] [Multimodal] Prediction of Cellular Morphology Changes under Perturbations with a Transcriptome-Guided Diffusion Model (Nature Communications 2025) [paper] [code] [中文解读] [ask deepwiki]

• [rBio-1] [Agent] [World Model] [Perturbation] rBio1-Training Scientific Reasoning LLMs with Biological World Models as Soft Verifiers (bioRxiv 2025) [preprint] [code] [中文解读] [ask deepwiki]

• [Scvi-hub] [Tool] [Dataset] Scvi-hub: An Actionable Repository for Model-Driven Single-Cell Analysis (Nature Methods 2025) [paper] [code] [中文解读] [ask deepwiki]

• [GraphVelo] [Dynamics] [Multimodal] [Gene Regulation] GraphVelo Allows for Accurate Inference of Multimodal Velocities and Molecular Mechanisms for Single Cells (Nature Communications 2025) [paper] [code] [中文解读] [ask deepwiki]

• [Stereo-Cell] [Spatial] [Dataset] [Multimodal] Stereo-Cell: Spatial Enhanced-Resolution Single-Cell Sequencing with High-Density DNA Nanoball-Patterned Arrays (Science 2025) [paper] [code] [中文解读] [ask deepwiki]

• [SToFM] [Spatial] [Foundation Model] [Representation Learning] SToFM: A Multi-scale Foundation Model for Spatial Transcriptomics (ICML 2025) [paper] [code] [中文解读] [ask deepwiki]

• [NicheFlow] [Spatial] [Dynamics] Modeling Microenvironment Trajectories on Spatial Transcriptomics with NicheFlow (NeurIPS 2025) [paper] [code]

• [scGPT-spatial] [Foundation Model] [Spatial] [Representation Learning] scGPT-spatial: Continual Pretraining of Single-Cell Foundation Model for Spatial Transcriptomics (bioRxiv 2025) [preprint] [code]

• [SpatialAgent] [Agent] [Spatial] [Multimodal] [Tool] SpatialAgent: An Autonomous AI Agent for Spatial Biology (bioRxiv 2025) [preprint] [code] [中文解读] [ask deepwiki]

• [CellFlux] [Perturbation] [Morphology] CellFlux: Simulating Cellular Morphology Changes via Flow Matching (ICML 2025) [paper] [code] [ask deepwiki]

• [CellCLIP] [Morphology] [Perturbation] [Multimodal] [Representation Learning] CellCLIP: Learning Perturbation Effects in Cell Painting via Text-Guided Contrastive Learning (NeurIPS 2025) [paper] [code]

• [CELTIC] [Morphology] [Protein] Cell context-dependent in silico organelle localization in label-free microscopy images (Nature Methods 2025) [paper] [code]

• [MorphoDiff] [Perturbation] [Morphology] MorphoDiff: Cellular Morphology Painting with Diffusion Models (ICLR 2025) [paper] [preprint] [code]

• [PRESCRIBE] [Perturbation] PRESCRIBE: Predicting Single-Cell Responses with Bayesian Estimation (NeurIPS 2025) [paper] [code]

• [GDE] [Representation Learning] [Morphology] [Related] Generative Distribution Embeddings: Lifting Autoencoders to the Space of Distributions for Multiscale Representation Learning (NeurIPS 2025) [paper] [preprint] [code]

• [CellPB] [Benchmark] [Perturbation] Benchmarking AI Models for in Silico Gene Perturbation of Cells (bioRxiv 2025) [preprint] [code] [ask deepwiki]

• [PerturBench] [Benchmark] [Perturbation] [Tool] Benchmarking Machine Learning Models for Cellular Perturbation Analysis (NeurIPS 2025) [paper] [code]

• [CellForge] [Agent] [Virtual Cell] [Perturbation] AI Scientists for Building Virtual-Cell Models (arXiv 2025) [preprint] [code] [中文解读] [ask deepwiki] [Research Square]

• [Cradle-VAE] [Perturbation] [Representation Learning] Cradle-VAE: Enhancing Single-Cell Gene Perturbation Modeling with Counterfactual Reasoning-based Artifact Disentanglement (AAAI 2025) [paper] [code]

• [XTransferCDR] [Perturbation] [Representation Learning] Learning Cross-Domain Representations for Transferable Drug Perturbations on Single-Cell Transcriptional Responses (AAAI 2025) [paper] [code]

• [Linear Perturbation Baselines] [Benchmark] [Perturbation] [Foundation Model] [Evaluation & Measurement] Deep-Learning-Based Gene Perturbation Effect Prediction Does Not Yet Outperform Simple Linear Baselines (Nature Methods 2025) [paper] [code] [ask deepwiki]

• [Drifting Islands] [Benchmark] [Representation Learning] [Evaluation & Measurement] Limitations of Cell Embedding Metrics Assessed Using Drifting Islands (Nature Biotechnology 2025) [paper] [code] [ask deepwiki]

• [GeneAgent] [Agent] [Gene Regulation] [Tool] GeneAgent: Self-Verification Language Agent for Gene-Set Analysis Using Domain Databases (Nature Methods 2025) [paper] [code] [ask deepwiki]

• [Cell Behavior Grammar] [Virtual Cell] [Review] [Tool] [Dynamics] Human Interpretable Grammar Encodes Multicellular Systems Biology Models to Democratize Virtual Cell Laboratories (Cell 2025) [paper] [code] [ask deepwiki]

• [GREmLN] [Foundation Model] [Gene Regulation] [Representation Learning] GREmLN: A Cellular Graph Structure Aware Transcriptomics Foundation Model (bioRxiv 2025) [preprint] [code] [中文解读] [ask deepwiki]

• [CausCell] [Perturbation] [Representation Learning] [Dynamics] Causal Disentanglement for Single-Cell Representations and Controllable Counterfactual Generation (Nature Communications 2025) [paper] [code] [中文解读] [ask deepwiki]

• [CLIP^n] [Morphology] [Perturbation] [Representation Learning] Transitive Prediction of Small-Molecule Function through Alignment of High-Content Screening Resources (Nature Biotechnology 2025) [paper] [code] [ask deepwiki]

• [DrugPT] [Perturbation] [Multimodal] DrugPT: A Flexible Framework for Integrating Gene and Chemical Representations in Perturbation Modeling (bioRxiv 2025) [preprint]

• [OmniPert] [Foundation Model] [Perturbation] OmniPert: A Deep Learning Foundation Model for Predicting Responses to Genetic and Chemical Perturbations in Single Cancer Cells (bioRxiv 2025) [preprint]

• [UNAGI] [Dynamics] [Perturbation] [Intervention Design] A Deep Generative Model for Deciphering Cellular Dynamics and in Silico Drug Discovery in Complex Diseases (Nature Biomedical Engineering 2025) [paper] [code] [ask deepwiki]

• [OmiCLIP] [Foundation Model] [Spatial] [Morphology] [Multimodal] A Visual-Omics Foundation Model to Bridge Histopathology with Spatial Transcriptomics (Nature Methods 2025) [paper] [code] [ask deepwiki]

• [OCTO-vc] [Virtual Cell] [Spatial] [Perturbation] [Morphology] OCTO-vc: Virtual Cells in Real Tissue (Noetik Technical Report 2025) [technical report] [online demonstration]

• [Cell-GraphCompass] [Foundation Model] [Gene Regulation] [Representation Learning] Cell-GraphCompass: Modeling Single Cells with Graph Structure Foundation Model (National Science Review 2025) [paper] [code] [ask deepwiki]

• [scPRINT] [Foundation Model] [Gene Regulation] [Representation Learning] scPRINT: Pre-training on 50 Million Cells Allows Robust Gene Network Predictions (Nature Communications 2025) [paper] [code] [ask deepwiki]

• [CellFM] [Foundation Model] [Representation Learning] [Perturbation] [Gene Regulation] CellFM: A Large-Scale Foundation Model Pre-trained on Transcriptomics of 100 Million Human Cells (Nature Communications 2025) [paper] [code] [中文解读] [ask deepwiki]

• [C2S-Scale] [Foundation Model] [Multimodal] [Representation Learning] [Perturbation] C2S-Scale: Scaling Large Language Models for Next-Generation Single-Cell Analysis (bioRxiv 2025) [preprint] [code] [中文解读] [ask deepwiki]

• [scNET] [Representation Learning] [Gene Regulation] scNET: Learning Context-Specific Gene and Cell Embeddings by Integrating Single-Cell Gene Expression Data with Protein-Protein Interactions (Nature Methods 2025) [paper] [code] [ask deepwiki]

• [ProCyon] [Foundation Model] [Protein] [Multimodal] ProCyon: A multimodal foundation model for protein phenotypes (bioRxiv 2025) [preprint] [code] [project]

• [SubCell] [Foundation Model] [Morphology] [Protein] [Representation Learning] SubCell: Proteome-aware vision foundation models for microscopy capture single-cell biology (bioRxiv 2025) [preprint] [code] [ask deepwiki]

• [Token-Mol 1.0] [Related] [Multimodal] Token-Mol 1.0: Tokenized Drug Design with Large Language Models (Nature Communications 2025) [paper] [code] [ask deepwiki]

• [Predictive Immunotherapy] [Virtual Cell] [Review] [Intervention Design] Virtual Cells for Predictive Immunotherapy (Nature Biotechnology 2025) [paper]

• [Predict, Explain, Discover] [Virtual Cell] [Review] [Perturbation] [Intervention Design] Virtual Cells: Predict, Explain, Discover (arXiv 2025) [preprint]

• [scTranslator] [Foundation Model] [Protein] [Multimodal] [Perturbation] A pre-trained large generative model for translating single-cell transcriptomes to proteomes (Nature Biomedical Engineering 2025) [paper] [preprint] [code]

• [MTIProteinImputation] [Protein] [Spatial] [Morphology] Imputing single-cell protein abundance in multiplex tissue imaging (Nature Communications 2025) [paper] [code]

• [Cell Maps] [Virtual Cell] [Multimodal] [Protein] [Morphology] [Dataset] Multimodal cell maps as a foundation for structural and functional genomics (Nature 2025) [paper] [code] [project]

• [CellFlow] [Perturbation] [Multimodal] CellFlow Enables Generative Single-Cell Phenotype Modeling with Flow Matching (bioRxiv 2025) [preprint] [code] [ask deepwiki]

• [Prophet] [Foundation Model] [Perturbation] [Multimodal] Scalable and universal prediction of cellular phenotypes enables in silico experiments (bioRxiv 2025) [preprint] [code] [ask deepwiki]

• [Ovarian Co-Culture Features] [Morphology] [Perturbation] [Benchmark] Evaluating Feature Extraction in Ovarian Cancer Cell Line Co-Cultures Using Deep Neural Networks (Communications Biology 2025) [paper] [code]

• [Grow AI Virtual Cells] [Virtual Cell] [Review] Grow AI Virtual Cells: Three Data Pillars and Closed-Loop Learning (Cell Research 2025) [paper] [中文解读]

• [PS] [Perturbation] [Tool] Decoding Heterogeneous Single-Cell Perturbation Responses (Nature Cell Biology 2025) [paper] [code] [ask deepwiki]

• [Mixscale] [Perturbation] [Gene Regulation] [Multimodal] [Dataset] Systematic Reconstruction of Molecular Pathway Signatures Using Scalable Single-Cell Perturbation Screens (Nature Cell Biology 2025) [paper] [code] [ask deepwiki]

• [scDrugMap] [Benchmark] [Foundation Model] [Perturbation] [Tool] scDrugMap: benchmarking large foundation models for drug response prediction (Nature Communications 2025) [paper] [code]

• [VCC Commentary] [Virtual Cell] [Benchmark] [Perturbation] [Review] Virtual Cell Challenge: Toward a Turing Test for the Virtual Cell (Cell 2025) [paper] [homepage] [beginner's guidance]

• [CZI Evaluation] [Benchmark] [Virtual Cell] [Review] [Evaluation & Measurement] Benchmarking and Evaluation of AI Models in Biology: Outcomes and Recommendations from the CZI Virtual Cells Workshop (arXiv 2025) [preprint] [中文解读]

• [GET] [Foundation Model] [Gene Regulation] [Multimodal] A Foundation Model of Transcription across Human Cell Types (Nature 2025) [paper] [code] [ask deepwiki]

🗓️ 2024 — 27 papers

• [OP3 Benchmark] [Benchmark] [Dataset] [Perturbation] A benchmark for prediction of transcriptomic responses to chemical perturbations across cell types (NeurIPS 2024) [paper] [code (project)] [NeurIPS 2023 competition]

• [biolord] [Perturbation] [Representation Learning] Disentanglement of single-cell data with biolord (Nature Biotechnology 2024) [paper] [preprint] [code]

• [scBERT Deeper Evaluation] [Evaluation & Measurement] [Foundation Model] [Benchmark] Deeper evaluation of a single-cell foundation model (Nature Machine Intelligence 2024) [paper] [preprint] [code]

• [Phenom-1] [Foundation Model] [Morphology] [Representation Learning] [Perturbation] Masked Autoencoders for Microscopy are Scalable Learners of Cellular Biology (CVPR 2024) [paper] [preprint] [code]

• [MolPhenix] [Morphology] [Multimodal] [Representation Learning] [Perturbation] How Molecules Impact Cells: Unlocking Contrastive PhenoMolecular Retrieval (NeurIPS 2024) [paper] [preprint]

• [scCello] [Foundation Model] [Representation Learning] Cell-ontology guided transcriptome foundation model (NeurIPS 2024) [paper] [preprint] [code]

• [AIDO.Cell] [Foundation Model] [Representation Learning] Scaling Dense Representations for Single Cell with Transcriptome-Scale Context (bioRxiv 2024) [preprint] [code]

• [scMulan] [Foundation Model] [Representation Learning] scMulan: A Multitask Generative Pre-Trained Language Model for Single-Cell Analysis (RECOMB 2024) [paper] [preprint] [code]

• [GPTCelltype] [Agent] [Evaluation & Measurement] Assessing GPT-4 for cell type annotation in single-cell RNA-seq analysis (Nature Methods 2024) [paper] [preprint] [code]

• [Zero-Shot Perturbation] [Foundation Model] [Perturbation] Efficient Fine-Tuning of Single-Cell Foundation Models Enables Zero-Shot Molecular Perturbation Prediction (arXiv 2024) [preprint]

• [TranSiGen] [Perturbation] [Representation Learning] [Intervention Design] Deep Representation Learning of Chemical-Induced Transcriptional Profile for Phenotype-Based Drug Discovery (Nature Communications 2024) [paper] [code] [ask deepwiki]

• [PRnet] [Perturbation] [Intervention Design] Predicting transcriptional responses to novel chemical perturbations using deep generative model for drug discovery (Nature Communications 2024) [paper] [code]

• [GenePT] [Foundation Model] [Representation Learning] Simple and Effective Embedding Model for Single-Cell Biology Built from ChatGPT (Nature Biomedical Engineering 2024) [paper] [code] [ask deepwiki]

• [SCimilarity] [Foundation Model] [Representation Learning] [Tool] A Cell Atlas Foundation Model for Scalable Search of Similar Human Cells (Nature 2024) [paper] [code] [ask deepwiki]

• [scFoundation] [Foundation Model] [Representation Learning] [Perturbation] [Gene Regulation] Large-Scale Foundation Model on Single-Cell Transcriptomics (Nature Methods 2024) [paper] [code] [ask deepwiki]

• [scGPT] [Foundation Model] [Representation Learning] [Multimodal] [Perturbation] [Gene Regulation] scGPT: Toward Building a Foundation Model for Single-Cell Multi-Omics Using Generative AI (Nature Methods 2024) [paper] [code] [ask deepwiki]

• [TamGen] [Related] [Multimodal] TamGen: Drug Design with Target-Aware Molecule Generation through a Chemical Language Model (Nature Communications 2024) [paper] [code] [ask deepwiki] [publication code archive]

• [GeneCompass] [Foundation Model] [Representation Learning] [Gene Regulation] [Perturbation] GeneCompass: Deciphering Universal Gene Regulatory Mechanisms with a Knowledge-Informed Cross-Species Foundation Model (Cell Research 2024) [paper] [code] [ask deepwiki]

• [scTab] [Foundation Model] [Representation Learning] scTab: Scaling Cross-Tissue Single-Cell Annotation Models (Nature Communications 2024) [paper] [code] [ask deepwiki]

• [SATURN] [Foundation Model] [Representation Learning] Toward Universal Cell Embeddings: Integrating Single-Cell RNA-Seq Datasets across Species with SATURN (Nature Methods 2024) [paper] [code] [ask deepwiki]

• [Cell2Sentence] [Foundation Model] [Multimodal] [Representation Learning] Cell2Sentence: Teaching Large Language Models the Language of Biology (ICML 2024) [paper] [code] [ask deepwiki]

• [LangCell] [Foundation Model] [Multimodal] [Representation Learning] LangCell: Language-Cell Pre-training for Cell Identity Understanding (ICML 2024) [paper] [code] [ask deepwiki]

• [CellPLM] [Foundation Model] [Representation Learning] [Spatial] CellPLM: Pre-training of Cell Language Model beyond Single Cells (ICLR 2024) [paper] [code] [ask deepwiki]

• [scPROTEIN] [Protein] [Representation Learning] scPROTEIN: A Versatile Deep Graph Contrastive Learning Framework for Single-Cell Proteomics Embedding (Nature Methods 2024) [paper] [code]

• [scLinear] [Protein] [Multimodal] scLinear Predicts Protein Abundance at Single-Cell Resolution (Communications Biology 2024) [paper] [code]

• [Perturbation Proteomics] [Protein] [Perturbation] [Review] AI-Empowered Perturbation Proteomics for Complex Biological Systems (Cell Genomics 2024) [paper]

• [Stanford PhD Thesis] [Virtual Cell] [Perturbation] [Intervention Design] [Review] Engineering Cells Using Artificial Intelligence (Stanford University PhD Thesis 2024) [paper] [GitHub Homepage] [Arc profile]

🧭 Browse the Repository

• Start Here
• Latest Updates
• Research Papers
• Overview Papers
• Datasets
• Challenges and Competitions
• Reports and Blogs
• Videos
• Historical and Foundational Works
• Related Resources

About, data, and maintenance

• Scope
• Inclusion Rules
• Repository & Data
• Contributing

📚 Overview Papers

Five high-signal overview and perspective papers are shown by default. Expand the rest only if you want broader background coverage.

• [Nature Review] Revisiting the blueprint for an interpretable virtual cell (Nature Reviews Genetics 2026) [paper]

• [Cell Review] A world model of the virtual cell (Cell 2026) [paper]

• [Nature Perspective] Towards Multimodal Foundation Models in Molecular Cell Biology (Nature 2025) [paper] [中文解读]

• [Nature Methods] The virtual cell (Nature Methods 2025) [paper]

• [Cell Perspective] How to Build the Virtual Cell with Artificial Intelligence: Priorities and Opportunities (Cell 2024) [paper] [中文解读]

More overview and perspective papers (13)

• [Nature News] Can AI Build a Virtual Cell? Scientists Race to Model Life's Smallest Unit (Nature 2025) [paper] [中文解读]

• [Nature] The Human Cell Atlas from a cell census to a unified foundation model (Nature 2024) [paper]

• [Nature Review] Interpretation, extrapolation and perturbation of single cells (Nature Reviews Genetics 2026) [paper]

• [npj Digital Medicine] AI-driven virtual cell models in preclinical research: technical pathways, validation mechanisms, and clinical translation potential (npj Digital Medicine 2025) [paper]

• [Nature Review] Adapting systems biology to address the complexity of human disease in the single-cell era (Nature Reviews Genetics 2025) [paper]

• [Nature Genetics] Causal machine learning for single-cell genomics (Nature Genetics 2025) [paper]

• [Nature Methods] Multimodal foundation transformer models for multiscale genomics (Nature Methods 2025) [paper]

• [Nature Methods] AI proteomics: from protein identification to virtual cells (Nature Methods 2026) [paper]

• [Review] AI virtual cells for drug discovery and pharmacology (Trends in Pharmacological Sciences 2026) [paper]

• [Cell Review] World models for biomedicine (Cell 2026) [paper]

• [Nature Methods] Towards predictive virtual embryos with genomics and AI (Nature Methods 2026) [paper]

• [Cell Perspective] Empowering Biomedical Discovery with AI Agents (Cell 2024) [paper] [中文解读]

• [Cell Review] Toward a Foundation Model of Causal Cell and Tissue Biology with a Perturbation Cell and Tissue Atlas (Cell 2024) [paper] [中文解读]

🗃️ Datasets

🧫 Perturbation and Cell-State Atlases

• [Arc Virtual Cell Atlas] Large-scale perturbation atlas and codebase from Arc Institute [resource] [repo]

• [Tahoe-100M] Tahoe-100M: Mapping drug-induced molecular phenotypes at single-cell resolution (Cell 2026) [paper] [preprint] [code]

• [X-Atlas/Orion] Genome-Wide Perturb-Seq Datasets via a Scalable Fix-Cryopreserve Platform for Training Dose-Dependent Biological Foundation Models (bioRxiv 2025) [paper] [dataset]

• [Primary CD4+ T cell Perturb-seq] Genome-scale perturb-seq in primary human CD4+ T cells maps context-specific regulators of T cell programs and human immune traits (Cell 2026) [paper] [dataset]

• [TRADE: HepG2 and Jurkat Perturb-seq] Transcriptome-wide analysis of differential expression in perturbation atlases (Nature Genetics 2025) [paper] [dataset]

More resources (11)

• [scPerturb] scPerturb: Harmonized Single-Cell Perturbation Data (Nature Methods 2024) [paper] [dataset] [code]

• [CIGS] High-Throughput Profiling of Chemical-Induced Gene Expression across 93,644 Perturbations (Nature Methods 2025) [paper] [dataset] [code]

• [L1000/CMap] A Next Generation Connectivity Map: L1000 Platform and the First 1,000,000 Profiles (Cell 2017) [paper] [dataset]

• [Sci-Plex] Massively Multiplex Chemical Transcriptomics at Single-Cell Resolution (Science 2019) [paper] [dataset] [code]

• [Perturb-seq datasets] Genome-scale and combinatorial Perturb-seq datasets from Replogle, Norman, Dixit, Adamson, and related screens [overview]

• [Virtual Cell Challenge] Community perturbation-prediction challenge and hidden evaluation resources [homepage]

• [iPSC CRISPRi Perturbation Atlas] A genome-scale CRISPRi perturbation atlas of human induced pluripotent stem cells: transcriptional and fitness phenotypes for 11,739 targeted genes in KOLF2.1J iPSCs (Nature Biotechnology 2026) [paper] [dataset] [code]

• [In vivo brain Perturb-seq] Genome-scale functional mapping of the mammalian whole brain with in vivo Perturb-seq: 1,947 disease-associated genes, over 7.7 million mouse brain cells (bioRxiv 2026) [preprint]

• [Human Cytokine Dictionary (Parse 10M PBMC)] A single-cell cytokine dictionary of human peripheral blood: 9.7 million PBMCs from 12 donors stimulated with 90 cytokines (bioRxiv 2025) [preprint] [dataset]

• [Immune Dictionary] Dictionary of immune responses to cytokines at single-cell resolution: 86 cytokines across more than 17 immune cell types in mouse lymph nodes in vivo (Nature 2023) [paper] [resource]

• [OP3] Open Problems perturbation-prediction dataset and benchmark: 146 compounds tested on human PBMCs, from the NeurIPS 2023 competition (NeurIPS 2024 Datasets and Benchmarks) [paper] [dataset] [code]

🔬 Single-Cell Reference Atlases

• [scBaseCount] scBaseCount: An AI agent-curated, standardized, auto-updated single-cell data repository (Cell 2026) [paper] [preprint] [code-scRecounter] [code-SRAgent]

• [CZ CELLxGENE Discover] A single-cell data platform for scalable exploration, analysis, and modeling of aggregated data (NAR 2025) [paper] [dataset]

• [Human Cell Atlas] International atlas of human cells and tissues [portal] [paper]

More resources (3)

• [Tabula Sapiens] A multiple-organ single-cell transcriptomic atlas of humans (Science 2022) [paper] [dataset]

• [Human BioMolecular Atlas Program] HuBMAP healthy human tissue atlas and common coordinate framework [portal] [paper]

• [GTEx] Genotype-Tissue Expression project for human tissue expression baselines [portal] [overview]

🖼️ Multimodal, Morphology, and Imaging

• [scGeneScope] scGeneScope: A Treatment-Matched Single Cell Imaging and Transcriptomics Dataset and Benchmark for Treatment Response Modeling (NeurIPS 2025 Datasets and Benchmarks Track) [paper] [dataset]

• [CPJUMP1] Three million images and morphological profiles of cells treated with matched chemical and genetic perturbations (Nature Methods 2024) [paper] [dataset] [code]

• [Cell Painting Gallery] Public high-content cell painting datasets from Broad and partners [dataset] [overview]

More resources (5)

• [RxRx] Recursion high-content cellular imaging datasets for perturbation and batch-correction research [datasets]

• [CM4AI] Cell Maps for Artificial Intelligence: AI-Ready Maps of Human Cell Architecture from Disease-Relevant Cell Lines (bioRxiv 2024) [paper] [dataset]

• [STAMP] Single-cell transcriptomics analysis and multimodal profiling through imaging (Cell 2025) [paper]

• [JUMP Cell Painting] JUMP Cell Painting dataset: morphological impact of 136,000 chemical and genetic perturbations, covering over 116,000 compounds, 12,602 overexpressed genes and 7,975 CRISPR knockouts in U2OS cells (bioRxiv 2023) [preprint] [dataset] [gallery]

• [NeurIPS 2021 Multimodal] A sandbox for prediction and integration of DNA, RNA, and proteins in single cells: multi-donor bone-marrow CITE-seq and 10x Multiome benchmark data (NeurIPS 2021 Datasets and Benchmarks) [paper] [dataset]

🗺️ Spatial and Tissue Context

• [HEST-1k] HEST-1k: A Dataset for Spatial Transcriptomics and Histology Image Analysis (NeurIPS 2024) [paper] [code]

• [TCGA virtual ST atlas] TCGA virtual spatial transcriptomics atlas: DeepSpot-M predicted transcriptome-wide ST for TCGA H&E (FF + FFPE; ~28.7k slides / 32 cancer types) (medRxiv 2026) [paper] [dataset] [code]

• [HEST Xenium virtual ST] HEST Xenium virtual spatial transcriptomics: DeepSpot-M predicted transcriptome-wide ST for 59 HEST-1k 10x Xenium samples (~13.3M cells) (medRxiv 2026) [paper] [dataset] [code]

More resources (5)

• [Spatial Perturb-seq] Spatial perturb-seq data for functional genomics within intact tissue architecture (Nature Communications 2026) [paper] [code]

• [Perturb-FISH] CRISPR screening with imaging-based spatial transcriptomics (Cell 2025) [paper]

• [SPAC-seq] Uncovering spatially resolved functional genomics with CRISPR screen sequencing: pooled CRISPR screens read out with whole-transcriptome spatial profiling in tumors (Cell 2026) [paper]

• [10x Genomics spatial datasets] Visium, Xenium, and related public spatial transcriptomics example datasets [datasets]

• [Vizgen MERFISH datasets] Public MERFISH example datasets for spatial transcriptomics [datasets]

🧬 Protein, Organelle, and Molecular Priors

• [Human Protein Atlas] Subcellular and tissue protein expression atlases [resource] [subcellular] [paper]

• [OpenCell] Endogenous protein tagging, localization, and interaction data for human cellular organization (Science 2022) [paper] [dataset]

• [ProtiCelli] Proteome-wide image generation resources for human cell protein localization (bioRxiv 2026) [paper] [code]

More resources (3)

• [SubCell] Proteome-aware microscopy foundation model resources based on HPA images (bioRxiv 2025) [paper] [code]

• [STRING] Protein functional association and interaction networks [resource] [paper]

• [OmniPath] Signaling, ligand-receptor, and causal network priors for multi-omics analysis [resource] [paper]

💊 Chemical, Drug, and Target Resources

• [Drug Repurposing Hub] Curated compound library with targets, mechanisms, and clinical annotations (Nature Medicine 2017) [paper] [resource]

• [DepMap/CCLE/PRISM] Cancer dependency, molecular profile, and drug-response resources for cell-line perturbation modeling [resource] [paper]

• [ChEMBL] Drug-like molecules, bioactivities, targets, and assays [resource] [paper]

More resources (2)

• [BindingDB] Protein-small molecule binding affinity knowledgebase [resource] [paper]

• [PubChem] Compound identifiers, structures, assays, and bioactivity records [resource] [paper]

🏆 Challenges and Competitions

Time-boxed competitions with hidden test sets and live leaderboards. Both entries below ask a version of the same question, whether a model can predict cell state under perturbation, at two different biological scales: cultured cell lines for the Arc challenge, a developing embryo for the NeurIPS one. Their current deadlines are mirrored at the top of News. Earlier community challenges lists the closed competitions whose datasets and lessons these build on.

🧪 Virtual Cell Challenge (Arc Institute)

The Virtual Cell Challenge (VCC) is Arc Institute's annual AI virtual cell competition, explicitly modeled on CASP: every edition generates a fresh perturbation benchmark, keeps the test split hidden, and runs a live leaderboard. Two editions have run so far. Note that virtualcellchallenge.org now serves the current 2026 edition, so the 2025 links below point to Arc's archived posts rather than to the challenge site.

1st Edition (2025)

• Question: can a computational model stand in for a real Perturb-seq experiment?

• Task: predict the single-cell transcriptional response to held-out CRISPRi gene knockdowns, in a cell context deliberately chosen to sit off the training distribution of existing foundation models.

• Data: a newly generated Perturb-seq dataset of about 300,000 H1 human embryonic stem cells (H1 hESC) covering 300 genetic perturbations, produced with 10x Genomics Flex chemistry and released in segments for training, validation, and testing.

• Metrics: Perturbation Discrimination Score (PDS), Differential Expression Score (DES), and Mean Absolute Error (MAE), implemented in Arc's cell-eval suite. A Generalist Prize introduced during the competition re-ranked the 50 top-scoring final entries on the average rank across seven metrics: the three challenge metrics plus four from the STATE model paper (Pearson delta, Spearman correlation of log fold change, AUPRC, and Spearman correlation of effect size).

• Timeline and scale: registration opened on 26 June 2025, winners were announced on 6 December 2025 at NeurIPS 2025. Over 5,000 people registered from 114 countries, over 1,200 teams submitted results, and over 300 teams made final submissions.

• Winners:

  • 1st place, $100k: BM_xTVC (BioMap Research), entry xTrimoSCPerturb, a hybrid of deep learning with classical statistical features, protein embeddings, and curated public data.
  • 2nd place, $50k: XLearning Lab (Sichuan University, Tianfu Jincheng Lab), a metric-driven conditional generation framework over pseudo-bulk representations.
  • 3rd place, $25k: Outlier (UChicago, Dartmouth, HKU), entry TransPert, cross-cell-line prediction from summary-level statistics with similarity-aware aggregation and global linear scaling.
  • Generalist Prize, $100k: Altos Labs (Institute of Computation), entry go-with-the-flow, a flow matching generative model that learns time-dependent dynamics directly in gene expression space.

• Reported takeaway: the winning entries all paired deep learning with classical statistical features. Arc's own wrap-up states that pure end-to-end neural networks did not yet outperform hybrid models. As of this entry none of the four winning entries has released code or a methods paper; the TransPert manuscript is described as forthcoming.

• [Launch] Arc Institute launches its inaugural virtual cell competition (Arc Institute 2025) [announcement]

• [Commentary] Virtual Cell Challenge: Toward a Turing Test for the Virtual Cell (Cell Commentary 2025) [paper]

• [Data] Behind the Data of the Virtual Cell Challenge (Arc Institute 2025) [post]

• [Results] Virtual Cell Challenge 2025 Wrap-Up: Winners and Reflections (Arc Institute 2025) [post]

• [Evaluation Code] cell-eval: the metrics suite used to score submissions [code]

• [Metric Analysis] Effects of Distance Metrics and Scaling on the Perturbation Discrimination Score (arXiv 2025) [paper]

2nd Edition (2026)

Round two is the edition currently hosted at virtualcellchallenge.org, sponsored by NVIDIA, 10x Genomics, and Ultima Genomics. The facts below are taken from the challenge homepage, its Evaluation page, and the accompanying Cell paper, as read on 13 October 2026.

• Question: does a perturbation model transfer to cell contexts in which it has never seen a perturbation?

• Task: multi-context generalization and zero-shot prediction. Participants predict how multiple cell lines respond to specified gene knockdowns, given only non-targeting control profiles for those lines. Predictions are scored against new experimental perturbation data that Arc generates from those cell lines and that no submitted model has trained on.

• Data: no net-new training dataset this year. Arc ran CRISPRi Perturb-seq with 10x Flex in six cell lines from different tissues. Participants receive only non-targeting control profiles (18,400 control cells per context: 46 guides × 400 cells) for three anonymized validation contexts (A–C), together with the 300 CRISPRi knockdowns to predict in each; the final test set uses a new 300-perturbation panel in three different contexts (D–F). Participants may train on the H1 hESC data from the 2025 edition plus any other public or private data.

• Submission format: one .vcc file (an AnnData wrapper built with the vcc CLI) covering all three contexts, 400 predicted cells per perturbation, raw counts over 18,533 genes; two scored submissions per team per day.

• Metrics: six metrics scored by cell-eval2 on its vcc2026 profile: perturbation discrimination (pds), expression accuracy (mse), DE log-fold-change accuracy (nmae), DE direction fidelity (fid), DE direction reach (reach), and DE significance overlap (jac); the perturbed gene itself is excluded. Each metric is reference-scaled within its context, with 0 = the cell-context mean baseline and 1 = a real replicate experiment (mean of five disjoint half-splits), so scores above 1 and below 0 are both possible. The overall score is the unweighted mean over six metrics × three contexts. Validation and final-test scores are not comparable, and final-test scores stay hidden until winners are announced. This is a substantial revision of the 2025 PDS/DES/MAE scheme.

• Submission policy: entries do not require open weights or code. Arc frames this as a deliberate choice to encourage industry participation alongside academia and individual entrants. It also makes this edition less reproducible than the 2025 one, so treat leaderboard positions as claims rather than as verifiable results.

• Timeline: registration opened 4 August 2026, submissions and the live leaderboard opened 20 August 2026, the final test set is released 22 October 2026, final submissions are due 5 November 2026, and winners are announced in late November 2026.

• Prizes: the top three models win prizes valued at $100,000, $50,000, and $25,000, each half cash and half NVIDIA Brev credits.

• [Homepage] Virtual Cell Challenge (Arc Institute 2026) [homepage] [Virtual Cell Initiative]

• [Paper] Virtual Cell Challenge 2026: Benchmarking zero-shot generalization across cellular contexts (Cell 2026) [paper]

• [Evaluation Code] cell-eval2: GPU-accelerated scorer and the vcc2026 metric specification [code] [metric spec]

• [Community Baseline] zero-shot-perturbation: an open, reproducible transfer baseline for the 2026 task (average the same knockdown's response across six public CRISPRi screens, restore the magnitude; validation rank 433, overall 0.137 in Oct 2026) with a leave-one-line-out benchmark scored by cell-eval2, a fast pseudobulk evaluator, and noise- and transfer-ceiling code that bounds what any transfer model can reach [code]

🐣 Virtual Embryo Challenge (NeurIPS 2026)

The Virtual Embryo Challenge is a NeurIPS 2026 Competition Track entry organized by the Qiu Lab at Stanford with Harvard, UC San Diego, MBZUAI, Carnegie Mellon, GenBio AI, and Vizgen. It moves the perturbation question from cultured cell lines into a developing embryo, and scores prediction across developmental time, 3D position, and gene knockout. The facts below are taken from the challenge page and the NeurIPS competition announcement as read on 21 August 2026.

• Question: can a generative model predict how an embryo takes shape across space, scale, time, and perturbation?

• Tasks: three, scored on shared hidden test sets. Temporal: predict gene-expression distributions at later developmental stages (train on E8.5 and E9.5, validate on E10.5, hidden test at E12.5). Spatial-temporal: jointly predict expression and 3D location across scales, scored separately on the heart and on the whole embryo. Perturbation: predict a held-out knockout from wild-type development plus one observed perturbation (Mab21l2 knockout for training, Gata4 for validation, β-catenin as the hidden test).

• Data: one MERFISH resource covering whole embryos and a heart-focused subset, about 1M cells across 11 time points from E6.75 to E12.5. The challenge page says "mammalian" without naming the species; the companion Virtual Embryo atlas is mouse, indexed by Theiler stage.

• Tracks: two, scored separately on identical hidden test sets. The Human Team follows a conventional ML competition workflow. The Agent Team requires the run to belong to the coding agent or LLM-driven system: a human may write the initial prompt, but may not inspect intermediate results and feed judgement back in. Because both tracks are scored on the same held-out data, the agent track is a direct comparison against human-designed pipelines.

• Metrics: still not named on the challenge page as of 13 October 2026. The page states only that both tracks are scored on the same metrics and hidden test sets; teams get no biological performance feedback before designating up to two official test submissions per task, which are then scored, posted to the public test leaderboard, and locked.

• Timeline: portal opened 10 August 2026, starter kit and reference baselines released 15 August 2026, test phase opens 27 October 2026 at 09:00 PDT (moved from 20 October; per the timeline page read on 27 October 2026), final submissions close 2 December 2026, official evaluation starts 4 December 2026, winners announced 11 December 2026 at NeurIPS.

• Prizes: $112K total from the Laude Institute Moonshots Seed Grant, split into $54K in winner prizes ($27K per track: one $8K first, two $5K second, and three $3K third prizes), an $8K Generality Award for the best single architecture that answers more than one task, $30K in travel awards, and a $20K Community Contribution Award.

• [Homepage] The Virtual Embryo Challenge: Generative Modeling of Embryogenesis Across Space, Scale and Time (NeurIPS 2026 Competition Track) [homepage] [atlas] [NeurIPS announcement]

• [Background] Towards Predictive Virtual Embryos with Genomics and AI (Nature Methods 2026) [paper]

🗂️ Earlier Community Challenges

Closed competitions that are still useful for their datasets, baselines, and lessons about evaluation. Most were hosted by Open Problems in Single-Cell Analysis, whose living benchmark is described in Luecken et al., Nature Biotechnology 2025.

• [NeurIPS 2023] Single-cell perturbation prediction: generalizing experimental interventions to unseen contexts (Open Problems · Kaggle 2023): predict small-molecule responses in held-out PBMC cell types; the resulting 146-compound dataset and benchmark became OP3. [event] [Kaggle] [benchmark paper] [dataset]

• [NeurIPS 2022] Multimodal Single-Cell Integration Across Time, Individuals, and Batches (Open Problems · Kaggle 2022): predict one modality from another in a 300,000-cell CD34+ hematopoietic stem and progenitor time course, with the test set taken from a later time point than any training data. [event] [Kaggle]

• [NeurIPS 2021] Multimodal Single-Cell Data Integration (Open Problems 2021): modality prediction, modality matching, and joint embedding on bone-marrow CITE-seq and Multiome data. [event] [results paper] [dataset paper]

• [Schmidt Center] Cancer Immunotherapy Grand Data Science Challenge (Broad Institute 2023): predict the effects of held-out T-cell gene knockouts from in vivo Perturb-seq and propose new perturbations; the Hacohen Lab then tested predictions in mouse models. [challenge] [results preprint]

• [Schmidt Center × CrunchDAO] Obesity ML Competition (Broad Institute 2025–2026): predict how CRISPR perturbations shift adipocytes between energy-storing and energy-burning states from single-cell data, in three rounds from December 2025 to July 2026; submissions are closed and top models are slated for lab testing. [challenge] [CrunchDAO]

📝 Reports and Blogs

• [Announcement] International, cross-sector collaboration commits nearly $2 billion to build foundational data for AI models to predict and treat disease: Biohub's Virtual Biology Initiative, with DOE, NIH, Google DeepMind, Isomorphic Labs and Meta (© by Biohub 2026) [announcement] [DOE release]

• [Symposium] AI Proteomics and Virtual Cell (© by Westlake University 2025) [media] [中文解读]

• [Report] Projections at the Frontier: Snapshot 2025 (© by Decoding Bio's Team 2025) [slide] [中文解读]

• [Post] Chan Zuckerberg Initiative's rBio Uses Virtual Cells to Train AI, Bypassing Lab Work (© by Michael Nuñez 2025) [blog]

More reports and blogs (5)

• [Blog] AI's Next Frontier: Modeling Life Itself (© by Chan Zuckerberg Initiative 2025) [blog] [中文解读]

• [Blog] The State of Research on Virtual Cell Modeling (© by Will Connell 2025) [blog]

• [Blog] What Are Virtual Cells? Learning "Universal Representations" of Life's Fundamental Unit (© by Elliot Hershberg 2025) [blog]

• [中文 Blog] 什么是虚拟细胞：AI 生物学的“登月时刻”和“苦涩教训” (© by 范阳 2025) [blog]

• [Introduction] Virtual Cells (© by Udara Jay 2025) [blog]

🎥 Videos

• [Arc Institute] Predicting Cellular Responses to Perturbation across Diverse Contexts with STATE [YouTube]

• [Valence Labs] Virtual Cells: Predict, Explain, Discover [YouTube]

• [EPFL] Virtual Cells and Digital Twins: AI in Personalized Medicine [YouTube]

More videos (4)

• [SciLifeLab] Emma Lundberg: AI Virtual Cells Could Revolutionize Biological Science [YouTube]

• [Chan Zuckerberg Initiative] AI Virtual Cell Models: How AI is Accelerating Science [YouTube]

• [Chan Zuckerberg Initiative] CZI's Vision for AI-Powered "Virtual Cells" [YouTube]

• [Podcast] Google DeepMind CEO: We Want to Build a Virtual Cell [YouTube]

🕰️ Historical and Foundational Works

• [HPA Cell Atlas] [Morphology] A subcellular map of the human proteome (Science 2017) [paper] [resource]

• [OpenCell] [Morphology] OpenCell: Endogenous tagging for the cartography of human cellular organization (Science 2022) [paper] [dataset]

• [Genome-scale Perturb-seq] [Perturbation] Mapping information-rich genotype-phenotype landscapes with genome-scale Perturb-seq (Cell 2022) [paper] [code]

• [GEARS] [Perturbation] Predicting transcriptional outcomes of novel multigene perturbations with GEARS (Nature Biotechnology 2023) [paper] [code] [ask deepwiki]

• [Geneformer] [Foundation Model] Transfer Learning Enables Predictions in Network Biology (Nature 2023) [paper] [code] [documentation]

• [CellOT] [Perturbation] Learning Single-Cell Perturbation Responses Using Neural Optimal Transport (Nature Methods 2023) [paper] [code] [ask deepwiki]

More historical and foundational works (23)

• [CPA] [Perturbation] Predicting cellular responses to complex perturbations in high-throughput screens (Molecular Systems Biology 2023) [paper] [code]

• [CellOracle] [Gene Regulation] [Perturbation] Dissecting cell identity via network inference and in silico gene perturbation (Nature 2023) [paper] [code]

• [chemCPA] [Perturbation] Predicting Cellular Responses to Novel Drug Perturbations at a Single-Cell Resolution (NeurIPS 2022) [paper] [code]

• [scBERT] [Foundation Model] scBERT as a large-scale pretrained deep language model for cell type annotation of single-cell RNA-seq data (Nature Machine Intelligence 2022) [paper] [code]

• [Review] [Perturbation] Machine learning for perturbational single-cell omics (Cell Systems 2021) [paper]

• [scGen] [Perturbation] scGen predicts single-cell perturbation responses (Nature Methods 2019) [paper] [code]

• [Norman 2019] [Perturbation] Exploring genetic interaction manifolds constructed from rich single-cell phenotypes (Science 2019) [paper] [dataset]

• [scVI] [Representation Learning] Deep generative modeling for single-cell transcriptomics (Nature Methods 2018) [paper] [code]

• [Perturb-seq 2016] [Perturbation] Perturb-Seq: Dissecting Molecular Circuits with Scalable Single-Cell RNA Profiling of Pooled Genetic Screens (Cell 2016) [paper] · companion: A Multiplexed Single-Cell CRISPR Screening Platform Enables Systematic Dissection of the Unfolded Protein Response (Cell 2016) [paper] [dataset]

• [tGPT] [Foundation Model] Generative Pretraining from Large-Scale Transcriptomes for Single-Cell Deciphering (iScience 2023) [paper] [code] [ask deepwiki]

• [Virtual Cell] Building the Next Generation of Virtual Cells to Understand Cellular Biology (Biophysical Journal 2023) [paper]

• [Research Highlight] [Virtual Cell] Simulating a Whole Cell (Nature Methods 2022) [paper]

• [sciPENN] [Protein] A Multi-Use Deep Learning Method for CITE-seq and Single-Cell RNA-seq Data Integration with Cell Surface Protein Prediction and Imputation (Nature Machine Intelligence 2022) [paper] [code]

• [totalVI] [Protein] Joint Probabilistic Modeling of Single-Cell Multi-Omic Data with totalVI (Nature Methods 2021) [paper] [code]

• [cTP-net] [Protein] Surface Protein Imputation from Single Cell Transcriptomes by Deep Neural Networks (Nature Communications 2020) [paper] [code]

• [CITE-seq] [Protein] Simultaneous Epitope and Transcriptome Measurement in Single Cells (Nature Methods 2017) [paper]

• [Comment] Personalized Medicine: Time for One-Person Trials (Nature Comment 2015) [paper]

• [Theory] [Virtual Cell] A Whole-Cell Computational Model Predicts Phenotype from Genotype (Cell 2012) [paper]

• [Virtual Cell] The Virtual Cell - A Candidate Co-Ordinator for "Middle-Out" Modelling of Biological Systems (BIB 2009) [paper]

• [VCell 7.7] [Virtual Cell] Virtual Cell Modelling and Simulation Software Environment (IET Systems Biology 2008) [paper] [software]

• Quantitative Cell Biology with the Virtual Cell (Trends in Cell Biology 2003) [paper]

• [Review] The Virtual Cell: A Software Environment for Computational Cell Biology (Trends in Biotechnology 2001) [paper]

• [Opinion] Whole-Cell Simulation: A Grand Challenge of the 21st Century (Trends in Biotechnology 2001) [paper]

🔗 Related Resources

• [Virtual Cell Challenge] Official challenge site for evaluation and community updates [homepage]

• [Virtual Embryo] Interactive mouse development atlas with 3D reconstructions, annotated sections, and open REST and MCP endpoints [site] [challenge]

• [Arc Virtual Cell Atlas] Large-scale perturbation atlas and codebase from Arc Institute [repo]

• [VCell Software] Long-running software environment for computational cell biology [site]

• [Noetik OCTO-vc] Technical report and demo for virtual cells in tissue [report] [demo]

🎯 Scope

Here, AIVC stands for Artificial Intelligence Virtual Cell, a term popularized by the Cell perspective "How to Build the Virtual Cell with Artificial Intelligence: Priorities and Opportunities." The repository is intentionally broad but practical: resources that help researchers understand, model, benchmark, or build virtual cells and related cellular foundation models.

• In scope: virtual cell perspectives, perturbation modeling, single-cell and multimodal foundation models, spatial and morphology modeling, biological AI agents, datasets, benchmarks, and community resources closely connected to virtual cell research.
• Also included: adjacent work that is broadly useful for the virtual cell community, especially when it contributes data, evaluation methods, or modeling tools for cellular systems.
• Usually out of scope: generic biomedical AI work with weak cell-modeling relevance, low-confidence secondary sources, broken links, or items that do not add clear value beyond more central references already listed here.

✅ Inclusion Rules

• Published papers come only from the journals and conferences listed in config/venue_policy.json: leading journals, the main tracks of leading ML conferences, and a short list of solid field journals. CI rejects any other venue.
• Preprints are included when they present a new method compared against established baselines, or a new dataset or benchmark. Reviews, perspectives and position papers are added once formally published.
• Technical reports from companies and institutes are included when they are first-party and release a model, dataset or code, or report a quantitative evaluation. A blog that announces a listed paper is linked from that paper. Substantive standalone blogs go to Reports and Blogs.
• Use primary-source links. Include code, datasets, project pages, or Chinese summaries when they are clearly useful.
• Keep entries concise and broadly reusable for readers who are scanning the field.
• Use the controlled multi-label taxonomy; assign tags from the actual task and evidence, not only model names.

🧰 Repository & Data

These links are mainly for reuse, contribution, and maintenance rather than day-to-day browsing.

• Searchable catalog — filter papers by keyword, year, topic, and publication status.
• Curation record — per-paper tagging rationale, source links and code-association checks.
• Structured paper data — machine-readable JSON backing the Research Papers section.
• Architecture — how structured data, generated views, validation, and automation fit together.
• Automation — how literature updates are proposed and validated.

🤝 Contributing

If you want to suggest a paper, dataset, benchmark, blog, or project, open an Issue or Pull Request. Please follow CONTRIBUTING.md for the submission format and quality bar.

⭐ Star History