Research map / Probabilistic and scientific modeling / Molecules and proteins
Needles in the Haystack: Addressing Signal Dilution Improves scRNA-seq Perturbation Response Modeling and Evaluation
Gabriel Mateo Mejia, Henry E Miller, Francis J.A. Leblanc, BO WANG, Brendan Swain, Lucas Paulo de Lima Camillo
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| train with WMSE instead of MSE (ablation) | GEARS · Replogle22 unseen gene prediction | fraction of ground-truth variance captured (%): 14.51 -> 40.56 | Improved |
| train with WMSE instead of MSE (ablation) | GEARS · Replogle22 unseen gene prediction | Pearson(delta) DEGs: t = 10.66 | Improved |
| train with WMSE instead of MSE (ablation) | GEARS · Replogle22 unseen gene prediction | NIR: t = 10.25 | Improved |
| train with WMSE instead of MSE (ablation) | GEARS · Norman19 combination prediction | WMSE and weighted delta R2 | Improved |
| train with WMSE instead of MSE (ablation) | FMLP, scLambda, scGPT · Replogle22 unseen gene prediction | multiple evaluation metrics | Improved |
| evaluate with WMSE instead of MSE | mean baseline vs DEG-corrected mean · Replogle22 | fold error reduction: 1.14 -> 3.26 | Improved |
| evaluate with WMSE instead of MSE | mean baseline vs DEG-corrected mean · Norman19 | fold error reduction: 1.65 -> 9.65 | Improved |
| increase probability of perturbing a gene (delta) | mean baseline · simulation | MSE: correlation 0.42 | Hurt |
| increase control bias beta | mean baseline · simulation | Pearson(delta): correlation 0.54 | Mixed |
| increase number of cells per perturbation np | mean baseline · simulation | MSE: correlation -0.35 | Improved |
| add DEG weighting to MSE and R2 metrics | metric calibration · Replogle22 | baseline ordering | Improved |
Closest papers by meaning
- PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (NeurIPS 2025 · similarity 0.92)
- PertEval-scFM: Benchmarking Single-Cell Foundation Models for Perturbation Effect Prediction (ICML 2025 · similarity 0.89)
- PRESCRIBE: Predicting Single-Cell Responses with Bayesian Estimation (NeurIPS 2025 · similarity 0.87)
- A benchmark for prediction of transcriptomic responses to chemical perturbations across cell types (NeurIPS 2024 · similarity 0.86)
- Contextualizing biological perturbation experiments through language (ICLR 2025 · similarity 0.86)
- PerturbReason: A Knowledge-Grounded Benchmark and Framework for Cell-State–Conditioned Mechanistic Reasoning of Perturbation Effects (NeurIPS 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
