Article

Method Choice, Not Biology, Determines In Silico Perturbation Results: A Systematic Evaluation of Eight Methods Across Four Datasets

2026-08-19

Abstract excerpt

Most in silico perturbation methods for single-cell transcriptomics have been validated only on individual datasets, leaving their reliability and generalizability unknown. Through systematic cross-method, cross-dataset benchmarking of eight methods spanning six mathematical frameworks across four datasets, we find that six of eight methods—including widely used VAE-based and tensor decomposition approaches—fail t...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: 278be97b-f840-5c13-b321-5d993d2b44e2
DOI: 10.64898/2026.08.11.744106

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Method Choice, Not Biology, Determines In Silico Perturbation Results: A Systematic Evaluation of Eight Methods Across Four DatasetsDOI 10.64898/2026.08.11.744106Read

PathwayBench: a multi-criterion benchmark of pseudobulk pathway activity scoring methods reveals rank-window competition as a mechanism of biological signal loss in single-cell RNA-seqSemantic proximity Read

Integrative, high-resolution analysis of single cell gene expression across experimental conditions with PARAFAC2-RISESemantic proximity Read

StickForStats: automated statistical assumption validation for reproducible computational biologySemantic proximity Read

Sparse autoencoders reveal organized biological knowledge but minimal regulatory logic in single-cell foundation models: a comparative atlas of Geneformer and scGPTSemantic proximity Read

Phenotypic reversion and target prioritization for cellular inflammation via representation learning with foundation modelsSemantic proximity Read

Extraction of biological signals by factorization enables the reliable analysis of single-cell transcriptomicsSemantic proximity Read

A novel Bayesian factor analysis method improves detection of genes and biological processes affected by perturbations in single-cell CRISPR screeningSemantic proximity Read

Multimodal foundation model predicts zero-shot functional perturbations and cell fate dynamicsSemantic proximity Read

A human cytokine response atlas to reconstruct underlying gene regulatory networksSemantic proximity Read

Signal, Bounds, and Baselines: Principles for Evaluating Virtual Cell Perturbation ModelsSemantic proximity Read

Three Classes of Confound in Gene-Regulatory-Network Inference: A Systematic Audit and Open-Source Diagnostic ToolkitSemantic proximity Read

Decomposing multi-scale dynamic regulation from single-cell multiomics with scMagnifySemantic proximity Read