Plausibility Is Not Prediction: Contrastive Evidence for LLM-Based Cellular Perturbation Reasoning

• URL: http://arxiv.org/abs/2606.01042v1
• Date: Sun, 31 May 2026 06:13:26 GMT
• Title: Plausibility Is Not Prediction: Contrastive Evidence for LLM-Based Cellular Perturbation Reasoning
• Abstract summary: Perturbation experiments are central to understanding cellular mechanisms.<n>They remain costly and sparse, motivating prediction of gene expression responses for unobserved conditions.<n>We introduce CORE, which reframes prediction as a comparison task by organizing evidence into positive and negative outcomes from related perturbations.
• Score: 27.902219493068824
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Perturbation experiments are central to understanding cellular mechanisms, but remain costly and sparse, motivating prediction of gene expression responses for unobserved conditions. A promising recent direction leverages large language models (LLMs) as "virtual cell" simulators-using stepwise, knowledge-grounded mechanistic reasoning to infer differential expression-pointing toward an interpretable, knowledge-driven paradigm that transcends purely data-driven approaches. However, we find that plausibility is not prediction: despite producing biologically plausible explanations, these methods fail to capture perturbation-specific effects: systematically overestimating differential expression, often underperforming a simple gene-frequency baseline in aggregate evaluations, and collapsing to chance-level performance at the per-gene level. This reveals a reliance on intrinsic gene response tendencies rather than true perturbation reasoning. We trace this failure to how evidence is presented: existing methods evaluate perturbation-gene pairs in isolation, without exposing how related perturbations differ in their effects on the same gene. To address this limitation, we introduce CORE (Contrastive Organization of Relational Evidence), which reframes prediction as a comparison task by organizing evidence into positive and negative outcomes from related perturbations. Using a biomedical knowledge graph for evidence retrieval, CORE improves calibration and substantially boosts perturbation-specific prediction in both LLM-based and non-LLM settings: for example, on drug-perturbation data, CORE-Reasoning improves Qwen3.5-9B aggregate metrics by up to 28.6%, while on generic perturbation data, CORE-Voting raises macro-per-gene AUROC from chance to 0.703 in average across four cell lines. This highlights contrastive evidence organization as essential to reliable LLM-based perturbation reasoning

Related papers

• Joint Model and Data Sparsification via the Marginal Likelihood [53.29070892356214]
  We propose jointly learning individual feature and sample relevancies, enabling simultaneous model and data sparsification.<n>This symmetric pruning of model and data offers a natural extension that preserves conjugacy.<n> Empirical results across diverse regression tasks affirm that a joint ARD approach consistently yields both sparse and robust prediction models.
  arXiv  Detail & Related papers  (2026-05-28T13:26:53Z)
• Learning Latent Dynamical Causal Processes for Single-Cell Perturbation Prediction [40.68914450179759]
  Single-cell perturbation prediction aims to infer how cells respond to unseen interventions.<n>We propose a latent dynamical causal generative model for single-cell perturbation data.<n>Under suitable conditions, the latent causal variables are recoverable up to standard equivalence classes.
  arXiv  Detail & Related papers  (2026-05-25T08:32:23Z)
• scDFM: Distributional Flow Matching Model for Robust Single-Cell Perturbation Prediction [12.48933770510505]
  We present scDFM, a generative framework based on conditional flow matching.<n> scDFM aligns perturbed and control populations beyond cell-level correspondences.
  arXiv  Detail & Related papers  (2026-02-06T17:00:21Z)
• Latent Causal Diffusions for Single-Cell Perturbation Modeling [83.47931153555321]
  We present a generative model that frames single-cell gene expression as a stationary diffusion process observed under measurement noise.<n> LCD outperforms established approaches in predicting the distributional shifts of unseen perturbation combinations in single-cell RNA-sequencing screens.<n>We develop an approach we call causal linearization via perturbation responses (CLIPR), which yields an approximation of the direct causal effects between all genes.
  arXiv  Detail & Related papers  (2026-01-20T16:15:38Z)
• Departures: Distributional Transport for Single-Cell Perturbation Prediction with Neural Schrödinger Bridges [51.83259180910313]
  A major bottleneck in gene function analysis is the unpaired nature of single-cell data.<n>We approximate Schrdinger Bridge (SB) to tackle unpaired single-cell perturbation data.<n>Our model effectively captures heterogeneous single-cell responses and achieves state-of-the-art performance.
  arXiv  Detail & Related papers  (2025-11-17T08:27:13Z)
• Unlasting: Unpaired Single-Cell Multi-Perturbation Estimation by Dual Conditional Diffusion Implicit Bridges [68.98973318553983]
  We propose a framework based on Dual Diffusion Implicit Bridges (DDIB) to learn the mapping between different data distributions.<n>We integrate gene regulatory network (GRN) information to propagate perturbation signals in a biologically meaningful way.<n>We also incorporate a masking mechanism to predict silent genes, improving the quality of generated profiles.
  arXiv  Detail & Related papers  (2025-06-26T09:05:38Z)
• TxPert: Leveraging Biochemical Relationships for Out-of-Distribution Transcriptomic Perturbation Prediction [11.083533122552396]
  We present TxPert, a new state-of-the-art method that leverages multiple biological knowledge networks to predict responses under OOD scenarios.<n>In particular, we present: (i) TxPert, a new state-of-the-art method that leverages multiple biological knowledge networks to predict responses under OOD scenarios; (ii) an in-depth analysis demonstrating the impact of graphs, model architecture, and data on performance; and (iii) an expanded benchmarking framework that strengthens evaluation standards for perturbation modeling.
  arXiv  Detail & Related papers  (2025-05-20T21:13:23Z)
• Predicting gene essentiality and drug response from perturbation screens in preclinical cancer models with LEAP: Layered Ensemble of Autoencoders and Predictors [4.882734501598445]
  Existing predictive models suffer from limited, generalisability and interpretability.<n>We introduce a framework of Layered Ensemble of Autoencoders and Predictors (LEAP)<n>LEAP consistently improves prediction performances in unscreened cell lines across modelling strategies.
  arXiv  Detail & Related papers  (2025-02-21T18:12:36Z)
• Generative Intervention Models for Causal Perturbation Modeling [80.72074987374141]
  In many applications, it is a priori unknown which mechanisms of a system are modified by an external perturbation.<n>We propose a generative intervention model (GIM) that learns to map these perturbation features to distributions over atomic interventions.
  arXiv  Detail & Related papers  (2024-11-21T10:37:57Z)
• Nonlinear Permuted Granger Causality [0.6526824510982799]
  Granger causal inference is a contentious but widespread method used in fields ranging from economics to neuroscience. To allow for out-of-sample comparison, a measure of functional connectivity is explicitly defined using permutations of the covariate set. Performance of the permutation method is compared to penalized variable selection, naive replacement, and omission techniques via simulation.
  arXiv  Detail & Related papers  (2023-08-11T16:44:16Z)
• CausalBench: A Large-scale Benchmark for Network Inference from Single-cell Perturbation Data [61.088705993848606]
  We introduce CausalBench, a benchmark suite for evaluating causal inference methods on real-world interventional data. CaulBench incorporates biologically-motivated performance metrics, including new distribution-based interventional metrics.
  arXiv  Detail & Related papers  (2022-10-31T13:04:07Z)
• Efficient Causal Inference from Combined Observational and Interventional Data through Causal Reductions [68.6505592770171]
  Unobserved confounding is one of the main challenges when estimating causal effects. We propose a novel causal reduction method that replaces an arbitrary number of possibly high-dimensional latent confounders. We propose a learning algorithm to estimate the parameterized reduced model jointly from observational and interventional data.
  arXiv  Detail & Related papers  (2021-03-08T14:29:07Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.