{"id":"302e2413-c971-46e3-8739-7bf70a5e9c81","arxiv_id":"2606.12838","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"OCOO-T is a flow-matching Transformer model that directly denoises continuous gene expression profiles to predict transcriptional responses to perturbations and reports state-of-the-art results on Tahoe100M, Replogle, and PBMC benchmarks.","lead":"The paper introduces OCOO-T, a minimalist flow-matching model built on a vanilla Transformer that predicts single-cell gene expression changes after genetic, chemical, or cytokine perturbations. A smart generalist might read it for a simpler, more scalable approach to virtual cell simulations that could aid drug discovery.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"No significant objection identified","rationale":"The reader's weakest assumption correctly isolates the architectural minimalism as the key untested element, but the provided abstract and placeholder full-text note do not supply evidence of a flaw in that assumption. The evaluations are presented as comprehensive, so the load-bearing risk is empirical rather than structural. No adjustment to UNVERDICTED is warranted without access to the actual results tables or ablations.","tokens_in":1744,"tokens_out":238,"duration_ms":12051,"concrete_test":"Reproduce the Tahoe100M evaluation using the exact patching/depatching procedure and conditioning described in the methods; verify that the reported metrics hold when the adaptive layer norm is replaced by standard layer norm while keeping all other components fixed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that a vanilla Transformer with adaptive layer normalization and in-context tokens suffices for SOTA perturbation response prediction—rests on empirical results rather than unstated assumptions about biological priors. The abstract's description of the architecture and benchmarks does not reveal an internal inconsistency or missing derivation that would invalidate the scaling or performance claims.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper introduces OCOO-T, a minimalist flow-matching-based virtual cell model that uses a vanilla Transformer operating directly on continuous gene expression profiles to predict single-cell transcriptional responses to genetic, chemical, and cytokine perturbations. Perturbation embeddings, dosage, and cell specificity are incorporated via adaptive layer normalization and in-context tokens, with patching/depatching for scalability to long profiles. The central claim is that this simple architecture achieves state-of-the-art performance on the Tahoe100M, Replogle, and PBMC benchmarks across diverse perturbations and cell types.","tokens_in":1793,"tokens_out":290,"duration_ms":13133,"significance":"If the performance claims hold with proper validation, this would be significant for AIVC modeling by showing that standard flow-matching and Transformer components can suffice without auxiliary encoders, hierarchical VAEs, or gene-interaction priors, potentially improving scalability and reproducibility. The emphasis on a parameter-light design using established techniques is a strength for the field.","major_comments":[{"comment":"Abstract: the assertion of state-of-the-art performance on Tahoe100M, Replogle, and PBMC benchmarks provides no quantitative metrics, baseline details, error analysis, or statistical comparisons, which is load-bearing for the central empirical claim and prevents verification of the reported improvements.","section":"Abstract"}],"minor_comments":[],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for their review and constructive comment. We address the concern about the abstract below and will incorporate the suggested changes in the revised manuscript.","responses":[{"response":"We agree that the abstract would be strengthened by including key quantitative metrics to support the SOTA claim. In the revised version, we will add concise performance highlights (e.g., primary metrics and baseline comparisons on each benchmark) drawn directly from the results tables, while preserving the abstract's length and readability. This addresses the verification concern without altering the manuscript's core claims.","revision_made":"yes","referee_comment":"[Abstract] Abstract: the assertion of state-of-the-art performance on Tahoe100M, Replogle, and PBMC benchmarks provides no quantitative metrics, baseline details, error analysis, or statistical comparisons, which is load-bearing for the central empirical claim and prevents verification of the reported improvements."}],"tokens_in":1319,"tokens_out":206,"duration_ms":8237,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The paper's main move is to strip away the auxiliary encoders, hierarchical VAEs, and gene-interaction priors that most prior work uses, and instead run a standard Transformer directly on patched expression profiles under a flow-matching objective. Perturbation, dosage, and cell-type information are injected only via adaptive layer norm and in-context tokens. That is the actual novelty: a deliberately simple architecture that still claims to scale to long transcriptional profiles and beat existing methods on Tahoe100M, Replogle, and PBMC.\n\nWhat it does cleanly is state the design choice up front and tie it to the scaling argument via patching and depatching. If the empirical results hold, the reduced complexity could matter for people who want something easier to train and extend than the current crop of AIVC models.\n\nThe soft spot is obvious and not minor: the abstract asserts state-of-the-art performance across diverse perturbations and cell types but gives no metrics, no baseline descriptions, no error bars, and no evaluation protocol. Without those, the central performance claim cannot be assessed, and the weakest assumption—that adaptive layer norm plus in-context tokens are enough without biological priors—remains untested in the supplied text. The stress-test note is right that there is no internal contradiction or missing derivation, but that only means the claim is purely empirical; the evidence for it is simply absent here.\n\nThis is the kind of paper a computational biology reading group might discuss if the full manuscript contains the missing numbers and a reproducible setup. It is worth a serious referee's time because the task is central to drug-discovery modeling and the minimalist framing is worth pressure-testing, even if heavy revision will be needed to make the results verifiable.","headline":"OCOO-T is a minimalist vanilla Transformer with flow-matching for perturbation prediction that claims SOTA on three benchmarks, but the abstract supplies zero numbers or protocol details to check the claim.","tokens_in":2254,"tokens_out":424,"would_cite":false,"duration_ms":15260,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"A vanilla Transformer with flow-matching and adaptive normalization predicts single-cell transcriptional responses to perturbations at state-of-the-art accuracy.","keywords":["virtual cell","transcriptional perturbation","flow matching","transformer","single-cell omics","perturbation response","drug discovery","gene expression prediction"],"falsifier":"A new benchmark dataset containing strong, previously unseen gene-regulatory interactions where any method that explicitly encodes those interactions significantly outperforms OCOO-T on held-out perturbations.","tokens_in":2650,"feed_emoji":"🧬","tokens_out":679,"duration_ms":14936,"temperature":0.7,"pith_summary":"The paper establishes that a minimalist flow-matching model built on a standard Transformer stack can predict how cells change their gene expression after genetic, chemical, or cytokine perturbations. It does so by treating the response as a continuous denoising process and injecting perturbation type, dosage, and cell identity only through adaptive layer normalization plus in-context tokens. This approach avoids the auxiliary encoders, hierarchical VAEs, or gene-interaction graphs used in prior work. A sympathetic reader would care because simpler architectures could make large-scale virtual-cell simulations practical for drug discovery and regulatory-network inference. Evaluations on Tahoe100M, Replogle, and PBMC data show the model matches or exceeds existing methods while scaling to long expression profiles via patching.","feed_headline":"Vanilla Transformer reaches SOTA on cell perturbation prediction","feed_subtitle":"OCOO-T conditions a standard denoising model with layer normalization and tokens to handle long gene profiles without extra encoders.","key_machinery":"Vanilla Transformer stack performing flow-matching denoising on continuous gene-expression profiles, conditioned by adaptive layer normalization and in-context tokens.","core_discovery":"OCOO-T formulates transcriptional perturbation response prediction as a continuous-time flow-matching denoising task performed by a vanilla Transformer that operates directly on continuous gene-expression vectors; perturbation embeddings, dosage, and cell specificity are supplied solely through adaptive layer normalization and in-context tokens, enabling state-of-the-art accuracy across diverse perturbations and cell types on Tahoe100M, Replogle, and PBMC benchmarks together with linear scaling to long profiles through patching and depatching.","pith_inferences":["If the minimalist conditioning proves sufficient, explicit gene-interaction graphs may be unnecessary for many perturbation-prediction tasks.","The same patching strategy could be tested on other high-dimensional single-cell modalities such as chromatin accessibility or protein abundance.","Training cost and iteration speed for virtual-cell models would drop if the vanilla-Transformer baseline continues to match specialized architectures."],"forward_implications":["The model scales linearly to full-length transcriptional profiles by patching and depatching cellular contexts.","Performance remains competitive across genetic, chemical, and cytokine perturbations as well as multiple cell types.","Architectural complexity can be reduced while preserving or improving accuracy on existing single-cell perturbation benchmarks.","In-silico cellular simulation becomes feasible at larger scale because the design avoids dedicated encoder-decoder modules."],"fun_headline_variants":["Transformer Predicts Perturbation Responses Directly","Vanilla Transformer Scales to Long Gene Profiles","Flow Matching with Standard Transformer on Cell Data","OCOO-T Conditions Denoising Model with Layer Norm","Patching Scales Transcriptional Perturbation Modeling"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"That perturbation type, dosage, and cell identity supplied only through adaptive layer normalization and in-context tokens are sufficient to capture relevant biological response dynamics without gene-interaction priors or hierarchical encoders.","fun_headline_variants_meta":{"raw":{"variants":["Transformer Predicts Perturbation Responses Directly","Vanilla Transformer Scales to Long Gene Profiles","Flow Matching with Standard Transformer on Cell Data","OCOO-T Conditions Denoising Model with Layer Norm","Patching Scales Transcriptional Perturbation Modeling"]},"model":"grok-4.3","cost_usd":0.008664,"raw_usage":{"total_tokens":3920,"prompt_tokens":694,"num_sources_used":0,"completion_tokens":65,"cost_in_usd_ticks":86637000,"prompt_tokens_details":{"text_tokens":694,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":3161,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":694,"tokens_out":65,"duration_ms":18419,"temperature":1.0,"reasoning_tokens":3161,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-27T05:20:07.440192+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A new benchmark dataset containing strong, previously unseen gene-regulatory interactions where any method that explicitly encodes those interactions significantly outperforms OCOO-T on held-out perturbations.","supporting_citations":[],"review_version":1}