{ "readiness_schema": "nmd-vcell-vcc-readiness/1.0", "feature_build": "VCC-READY-20260729-02", "assessed_at": "2026-07-29", "challenge_watch": { "organizer": "Arc Institute", "launch_date": "2026-08-20", "announcement_version": "official-organizer-social-announcement-2026-07", "checked_at": "2026-07-29", "announced_scope": "A new prediction problem, a wider range of metrics and a USD 100,000 grand prize; task details remain undisclosed.", "task_details_state": "NOT_YET_ANNOUNCED", "stale_after_days": 7, "stale_behavior": "Display the last checked date and retain the NOT_YET_ANNOUNCED state until an official task specification is verified.", "update_behavior": "Bind task, data, split and metric fields only from a versioned official source; preserve the prior announcement snapshot.", "source_url": "https://www.linkedin.com/posts/arc-institute-org_start-assembling-your-team-because-the-2026-activity-7477802662269186048-SXyD" }, "evidence_sources": [ { "source_id": "arc_vcc_2026_announcement", "title": "Arc Institute 2026 Virtual Cell Challenge announcement", "source_type": "official_organizer_announcement", "verification_state": "VERIFIED_PRIMARY", "evidence_grade": "A_authoritative_event_source", "url": "https://www.linkedin.com/posts/arc-institute-org_start-assembling-your-team-because-the-2026-activity-7477802662269186048-SXyD" }, { "source_id": "arc_vcc_2025_wrap", "title": "Virtual Cell Challenge 2025 Wrap-Up: Winners and Reflections", "source_type": "official_organizer_postmortem", "verification_state": "VERIFIED_PRIMARY", "evidence_grade": "A_authoritative_benchmark_source", "url": "https://arcinstitute.org/news/virtual-cell-challenge-2025-wrap-up" }, { "source_id": "arc_vcc_2025_data_metrics", "title": "Behind the Data of the Virtual Cell Challenge", "source_type": "official_organizer_metric_definition", "verification_state": "VERIFIED_PRIMARY", "evidence_grade": "A_authoritative_benchmark_source", "url": "https://arcinstitute.org/news/behind-the-data-virtual-cell-challenge" }, { "source_id": "arc_vcc_dataset_readme", "title": "Arc Virtual Cell Challenge dataset README", "source_type": "official_data_repository", "verification_state": "VERIFIED_PRIMARY", "evidence_grade": "A_reproducible_repository", "url": "https://github.com/ArcInstitute/arc-virtual-cell-atlas/blob/main/virtual-cell-challenge/README.md" }, { "source_id": "arc_cell_eval", "title": "cell-eval perturbation model evaluation suite", "source_type": "official_open_source_software", "verification_state": "VERIFIED_PRIMARY", "evidence_grade": "A_reproducible_repository", "url": "https://github.com/ArcInstitute/cell-eval" }, { "source_id": "primeflow_2026", "title": "PRiMeFlow: Capturing Complex Expression Heterogeneity in Perturbation Response Modelling", "source_type": "preprint", "verification_state": "VERIFIED_EXISTENCE_PREPRINT", "evidence_grade": "B_preprint_not_peer_reviewed", "url": "https://arxiv.org/abs/2604.13986" }, { "source_id": "lingshu_cell_2026", "title": "Lingshu-Cell: A generative cellular world model for transcriptome modeling toward virtual cells", "source_type": "preprint", "verification_state": "VERIFIED_EXISTENCE_PREPRINT", "evidence_grade": "B_preprint_not_peer_reviewed", "url": "https://arxiv.org/abs/2603.25240" }, { "source_id": "in_the_wild_benchmark_2026", "title": "Benchmarking virtual cell models for in-the-wild perturbation response", "source_type": "preprint", "verification_state": "VERIFIED_EXISTENCE_PREPRINT", "evidence_grade": "B_preprint_not_peer_reviewed", "url": "https://arxiv.org/abs/2604.27646" } ], "status_vocabulary": { "AVAILABLE": "Implemented and exposed in the frozen public benchmark.", "PARTIAL": "A related asset exists, but it is not task-equivalent or lacks a required context.", "MISSING": "Not implemented or not supported by the current frozen data." }, "layers": [ { "layer_id": "baseline", "label": "Baseline layer", "purpose": "Require complex models to beat simple, interpretable comparators.", "checks": [ { "check_id": "zero_change_control_mean", "label": "Zero-change / control-mean baseline", "status": "AVAILABLE", "evidence": "Repeated-fold RMSE comparison against zero response is frozen.", "next_action": "Retain as a mandatory comparator in every future task." }, { "check_id": "outer_train_mean", "label": "Outer-training response mean", "status": "AVAILABLE", "evidence": "Leakage-safe train-mean comparator is evaluated in every outer fold.", "next_action": "Keep fold-specific fitting and reporting." }, { "check_id": "regularized_linear", "label": "Regularized linear model", "status": "AVAILABLE", "evidence": "The current safe external model is a strongly shrunk linear/ridge estimator.", "next_action": "Preserve as the interpretable model floor." }, { "check_id": "pseudobulk", "label": "Target-level pseudo-bulk", "status": "AVAILABLE", "evidence": "Cells are aggregated to target-level response objects before inference.", "next_action": "Publish the exact aggregation recipe when source-level raw objects are available." }, { "check_id": "nearest_neighbor", "label": "Nearest-neighbour perturbation transfer", "status": "MISSING", "evidence": "No frozen nearest-neighbour comparator is registered.", "next_action": "Add training-only perturbation similarity and a leakage-safe neighbour baseline." }, { "check_id": "public_foundation_model", "label": "Public foundation-model comparator", "status": "MISSING", "evidence": "No public model output has been imported under the same split and metric contract.", "next_action": "Evaluate an openly reproducible model only after exact input and output parity is established." } ] }, { "layer_id": "hidden_generalization", "label": "Hidden generalization layer", "purpose": "Separate same-dataset interpolation from biologically meaningful transfer.", "checks": [ { "check_id": "same_context_unseen_target", "label": "Same-context unseen perturbation", "status": "AVAILABLE", "evidence": "Twenty deterministic repeated balanced five-fold realizations hold out entire HepG2 perturbation targets.", "next_action": "Keep this as the lowest generalization tier, not as disease transfer." }, { "check_id": "leave_one_donor_out", "label": "Leave-one-donor-out", "status": "MISSING", "evidence": "The current direct perturbation substrate does not expose multiple donors.", "next_action": "Acquire donor-resolved muscle/DMD perturbation data and lock donor-disjoint splits." }, { "check_id": "leave_one_dataset_out", "label": "Leave-one-dataset-out", "status": "MISSING", "evidence": "External muscle datasets have different roles and do not form a harmonized perturbation response panel.", "next_action": "Create accession-disjoint training and test objects after modality harmonization." }, { "check_id": "leave_one_disease_stage_out", "label": "Leave-one-disease-stage-out", "status": "MISSING", "evidence": "No stage-resolved perturbation outcome matrix is available.", "next_action": "Define stage labels and hold out complete disease stages, never random cells." }, { "check_id": "cross_cell_context", "label": "Cross-cell-type / context transfer", "status": "PARTIAL", "evidence": "A 55-target Frangieh external aggregate-delta diagnostic is executed and concludes NO_DIRECTIONAL_TRANSFER_SUPPORT; it is not muscle or DMD truth.", "next_action": "Generate or import matched muscle-context perturbation outcomes." }, { "check_id": "perturbed_timecourse", "label": "Perturbation-conditioned time course", "status": "PARTIAL", "evidence": "GSE52529 supplies unperturbed 0/24/48/72-hour myogenic reference states only.", "next_action": "Sample the same candidate perturbations across matched differentiation timepoints." } ] }, { "layer_id": "multi_metric", "label": "Multi-metric layer", "purpose": "Prevent a single metric from hiding magnitude, direction or distribution failures.", "checks": [ { "check_id": "global_expression_error", "label": "Global expression error", "status": "AVAILABLE", "evidence": "Repeated-fold RMSE is frozen; MAE-delta is now executed on the 55-target external diagnostic beside identical zero and train-mean comparators.", "next_action": "Rebind MAE only after the official 2026 expression scale and prediction unit are known." }, { "check_id": "direction_recovery", "label": "Perturbation direction recovery", "status": "AVAILABLE", "evidence": "Raw/residual cosine plus external delta Pearson and delta Spearman endpoints are reported; the external directional result remains negative.", "next_action": "Retain the current negative result: raw-direction support is mixed." }, { "check_id": "des_deg_recovery", "label": "DES / DEG recovery", "status": "MISSING", "evidence": "No VCC-compatible differential-expression score is computed.", "next_action": "Add up/down DEG precision-recall and threshold sensitivity using cell-eval-compatible outputs." }, { "check_id": "pds_discrimination", "label": "PDS / perturbation discrimination", "status": "PARTIAL", "evidence": "Raw L1 retrieval, a scale sweep and truth-norm-matched sensitivity are executed on 55 aggregate external targets; this is not an official AnnData or challenge run.", "next_action": "Re-execute with official 2026 AnnData inputs and preserve raw plus scale-audit outputs." }, { "check_id": "distribution_distance", "label": "Single-cell distribution distance", "status": "MISSING", "evidence": "The current model predicts target-level aggregates, not cell distributions.", "next_action": "Gate Wasserstein/MMD or cell-eval distribution metrics on genuine cell-level predictions." }, { "check_id": "pathway_recovery", "label": "Pathway / program recovery", "status": "PARTIAL", "evidence": "Frozen pathway panels are descriptive overlaps, not prediction-versus-truth metrics.", "next_action": "Score prespecified program changes against held-out observed responses." }, { "check_id": "cell_state_proportions", "label": "Cell-state proportion recovery", "status": "MISSING", "evidence": "No generative cell-distribution output or matched proportion truth is available.", "next_action": "Evaluate only when model outputs represent cell-level distributions." }, { "check_id": "uncertainty", "label": "Uncertainty and abstention", "status": "AVAILABLE", "evidence": "Hierarchical bootstrap intervals, strict gates and an abstention boundary are frozen.", "next_action": "Extend uncertainty to donor, dataset and disease-stage components." } ] }, { "layer_id": "disease_loop", "label": "Disease closure layer", "purpose": "Link expression predictions to independent NMD/DMD biological outcomes.", "checks": [ { "check_id": "dmd_observational_context", "label": "DMD observational context", "status": "AVAILABLE", "evidence": "Four frozen DMD evidence channels expose observed direction and source agreement.", "next_action": "Use as context qualification, never as perturbation truth." }, { "check_id": "human_myoblast_function", "label": "Human-myoblast perturbation phenotype", "status": "PARTIAL", "evidence": "GSE293514 provides bounded KO/fusion evidence for 9 of 21 candidates.", "next_action": "Expand candidate coverage and measure multiple muscle functions." }, { "check_id": "dmd_correction_transcriptome", "label": "DMD-correction transcriptome", "status": "PARTIAL", "evidence": "GSE272233 is an orthogonal correction reference; candidates were not perturbed.", "next_action": "Use as a disease-state benchmark, not as candidate causal evidence." }, { "check_id": "prospective_registry", "label": "Prospective prediction registry", "status": "PARTIAL", "evidence": "Schemas and an append-only outcome board exist, but contain zero registered outcomes.", "next_action": "Register predictions before assays and bind them to immutable outcome definitions." }, { "check_id": "functional_endpoints", "label": "Regeneration, fibrosis, inflammation and vascular endpoints", "status": "MISSING", "evidence": "No unified candidate perturbation outcome panel measures these disease functions.", "next_action": "Prespecify one functional primary endpoint and bounded secondary endpoints per study." }, { "check_id": "clinical_translation", "label": "Motor function / drug or clinical response", "status": "MISSING", "evidence": "No validated link from current model outputs to clinical outcomes exists.", "next_action": "Keep clinical utility locked until independent prospective validation." } ] } ], "claim_boundary": "This readiness matrix is an implementation audit and competition-preparation plan. It does not claim VCC 2026 task compatibility before the rules are released, nor does it convert same-HepG2 benchmark support into muscle, DMD, treatment or clinical validity.", "status_counts": { "total": 26, "available": 9, "partial": 7, "missing": 10 }, "api": "api/v1.1/vcc_readiness.json", "human_readable_audit": "downloads/VCC_2026_BENCHMARK_READINESS.md" }