Protocol

PerturBench v4 mode-collapse diagnostics on Frangieh21

Mode-collapse diagnostics on the Frangieh21 covariate-transfer split: RMSE, RMSE rank, transposed rank and matrix distance between predicted and observed cosine similarity matrices.

5 evaluations · 20 results

Overview

An explanatory profile has not yet been reviewed for this record.

Consult the linked sources for architecture or protocol details. Missing evidence is not evidence of a missing capability.

5 recorded evaluations, 20 metric rows. A comparison chart has not yet been validated for these results. The table retains the individual findings and their sources.

View coverage and remaining gaps across all benchmarks

Results

Results are available, but no reviewed comparison panel is linked in this release.

All evaluations

5 evaluations · 20 results. Different protocols are not a single leaderboard.

Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 · Dataset subset: Frangieh21 (PerturBench v4 co-culture split)

Sorted by Perturbench matrix distance (lower is better). The best value in each column is highlighted. Decimals are rounded for display; each value links to the printed value and its source.

Tested configuration	Perturbench matrix distance	Perturbench RMSE of mean	Perturbench RMSE rank	Perturbench RMSE transposed rank
SAMS-VAE* (S)Evaluation	54.5 (best in column)	0.023 (best in column)	0.16	0.416
LAEvaluation	65.3	0.023 (best in column)	0.127 (best in column)	0.413 (best in column)
CPA* (noAdv)Evaluation	80.2	0.024	0.156	0.48
CPA*Evaluation	80.4	0.024	0.236	0.484
SAMS-VAE*Evaluation	81	0.024	0.383	0.491

All 20 result rows with coverage, uncertainty and sources

Filter evaluations

Tested model or methodProtocolDatasetEvaluation setupMetricEvidence origin

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configuration	Protocol and dataset	Finding	Evidence and details
Configuration: CPA*	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	80.4 Perturbench matrix distance unit-unreported · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source CPA ∗ on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'CPA ∗', column 'Matrix distance'
Configuration: CPA* (noAdv)	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	80.2 Perturbench matrix distance unit-unreported · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CPA ∗ (noAdv) on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'CPA ∗ (noAdv)', column 'Matrix distance'
Configuration: CPA* (noAdv)	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.156 Perturbench RMSE rank fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CPA ∗ (noAdv) on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'CPA ∗ (noAdv)', column 'RMSE mean rank'
Configuration: CPA* (noAdv)	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.024 Perturbench RMSE of mean unit-unreported · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CPA ∗ (noAdv) on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'CPA ∗ (noAdv)', column 'RMSE mean'
Configuration: CPA* (noAdv)	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.48 Perturbench RMSE transposed rank fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source CPA ∗ (noAdv) on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'CPA ∗ (noAdv)', column 'RMSE mean transposed-rank'
Configuration: CPA*	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.236 Perturbench RMSE rank fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source CPA ∗ on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'CPA ∗', column 'RMSE mean rank'
Configuration: CPA*	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.024 Perturbench RMSE of mean unit-unreported · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source CPA ∗ on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'CPA ∗', column 'RMSE mean'
Configuration: CPA*	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.484 Perturbench RMSE transposed rank fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source CPA ∗ on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'CPA ∗', column 'RMSE mean transposed-rank'
Configuration: LA	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	65.3 Perturbench matrix distance unit-unreported · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LA on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'LA', column 'Matrix distance'
Configuration: LA	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.127 Perturbench RMSE rank fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LA on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'LA', column 'RMSE mean rank'
Configuration: LA	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.023 Perturbench RMSE of mean unit-unreported · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LA on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'LA', column 'RMSE mean'
Configuration: LA	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.413 Perturbench RMSE transposed rank fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source LA on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'LA', column 'RMSE mean transposed-rank'
Configuration: SAMS-VAE*	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	81 Perturbench matrix distance unit-unreported · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source SAMS-VAE ∗ on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'SAMS-VAE ∗', column 'Matrix distance'
Configuration: SAMS-VAE*	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.383 Perturbench RMSE rank fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source SAMS-VAE ∗ on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'SAMS-VAE ∗', column 'RMSE mean rank'
Configuration: SAMS-VAE*	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.024 Perturbench RMSE of mean unit-unreported · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source SAMS-VAE ∗ on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'SAMS-VAE ∗', column 'RMSE mean'
Configuration: SAMS-VAE*	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.491 Perturbench RMSE transposed rank fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source SAMS-VAE ∗ on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'SAMS-VAE ∗', column 'RMSE mean transposed-rank'
Configuration: SAMS-VAE* (S)	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	54.5 Perturbench matrix distance unit-unreported · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source SAMS-VAE ∗ (S) on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'SAMS-VAE ∗ (S)', column 'Matrix distance'
Configuration: SAMS-VAE* (S)	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.16 Perturbench RMSE rank fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source SAMS-VAE ∗ (S) on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'SAMS-VAE ∗ (S)', column 'RMSE mean rank'
Configuration: SAMS-VAE* (S)	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.023 Perturbench RMSE of mean unit-unreported · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source SAMS-VAE ∗ (S) on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'SAMS-VAE ∗ (S)', column 'RMSE mean'
Configuration: SAMS-VAE* (S)	Protocol: PerturBench v4 mode-collapse diagnostics on Frangieh21 Dataset subset: Frangieh21 (PerturBench v4 co-culture split)	0.416 Perturbench RMSE transposed rank fraction · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source SAMS-VAE ∗ (S) on PerturBench v4 frangieh21-collapse lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse Aggregation: Not reported PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Table 13, row 'SAMS-VAE ∗ (S)', column 'RMSE mean transposed-rank'

Source checking is not independent reproduction. Release 2026-10-23-5347695500ed.

Methods and evaluation design

Procedure, tasks and evaluated configurations

Recorded evaluations

Each evaluation records what was tested and under which conditions.

Baseline coverage

Reference methods help show what a model adds beyond simple controls. We track a null control and a conventional method for each protocol.

0 of 2 active baseline roles have published Rewire measurements in this release. Measurements on a selected protocol do not establish coverage of an entire suite.

No execution recipe linked to this protocol. Recipe availability does not establish a completed evaluation.

Author-reported evaluations: 3
External evaluations: 2

Literature evidence is not a Rewire measurement. Executed but unpublished runs and private review status are not included.

Null control

Proposed control: requires review

No-change prediction under matched control conditions

Protocol-specific applicability, permitted inputs, access, split, evaluator and execution requirements need review before implementation or execution.

This is a suggested selection rule, not a validated method or a measured score.

Conventional reference

Proposed control: requires review

Training-only mean-effect or linear prediction

Protocol-specific applicability, permitted inputs, access, split, evaluator and execution requirements need review before implementation or execution.

This is a suggested selection rule, not a validated method or a measured score.

Protocol coverage CSV (gzip) · Model evaluation matrix (gzip) · Source table (gzip) · Release and checksums (gzip)

Coverage is derived from release 2026-10-23-5347695500ed. Source citations describe the original records; they do not validate an unreviewed baseline proposal. No results have been generated by this audit.

Run instructions

No runnable recipe has been reviewed for this protocol. Dataset access, model requirements, licences and compute requirements must be checked against its sources before execution.

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.

Evidence table

Inspect claims, sources and review details

Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.

One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.

0 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-10-23-5347695500ed
Property and statement	Original source and location	Review and provenance

No evidence rows match these filters. Choose another scope or clear the search.

How to read the evidence

Sources and history

Release 2026-10-23-5347695500ed · Record review: source checked

1 source record and release history

PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis (arXiv v4) · Original source · arXiv 2408.10609v4, posted 2025-10-24 (NeurIPS 2025 Datasets and Benchmarks Track version)

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: lit-cells-networks-20261011-perturbench-v4-protocol-frangieh21-mode-collapse

areas: single-cell
contexts: research
protocol: Mode-collapse diagnostics in Appendix C.7. Single values without spread; the number of runs behind each value is not printed.
version: PerturBench arXiv v4
metric: perturbench-matrix-distance
metric direction: lower
unit: unit-unreported
limitations: No spread or seed count is printed for these tables; values differ from the 5-seed means for the same models in Tables 5 and 8.
source locator: Table 13; Appendix C.7.3

Related records

part of: PerturBench
uses data: Frangieh21 (PerturBench v4 co-culture split)
assessment: CPA ∗ on PerturBench v4 frangieh21-collapse
assessment: CPA ∗ (noAdv) on PerturBench v4 frangieh21-collapse
assessment: LA on PerturBench v4 frangieh21-collapse
assessment: SAMS-VAE ∗ on PerturBench v4 frangieh21-collapse
assessment: SAMS-VAE ∗ (S) on PerturBench v4 frangieh21-collapse

Suggest a correction