Configuration

scGPT (Csendes et al. 2025)

Pretrained scGPT fine-tuned on each benchmark dataset, as reproduced by the authors.

Overview

Pretrained scGPT fine-tuned on each benchmark dataset, as reproduced by the authors.

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

Evaluations and results

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

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: scGPT (Csendes et al. 2025)	Protocol: Adamson et al. 2016 Perturb-seq, K562 CRISPRi single perturbations: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Adamson et al. 2016 Perturb-seq, K562 CRISPRi single perturbations (GEARS processing, Csendes et al. 2025)	0.99 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Adamson (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-adamson-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell C15, row 'Adamson / scGPT', column 'Pearson'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Adamson et al. 2016 Perturb-seq, K562 CRISPRi single perturbations: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Adamson et al. 2016 Perturb-seq, K562 CRISPRi single perturbations (GEARS processing, Csendes et al. 2025)	0.979 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Adamson (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-adamson-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell D15, row 'Adamson / scGPT', column 'Pearson DE'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Adamson et al. 2016 Perturb-seq, K562 CRISPRi single perturbations: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Adamson et al. 2016 Perturb-seq, K562 CRISPRi single perturbations (GEARS processing, Csendes et al. 2025)	0.641 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Adamson (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-adamson-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell E15, row 'Adamson / scGPT', column 'Pearson Delta'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Adamson et al. 2016 Perturb-seq, K562 CRISPRi single perturbations: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Adamson et al. 2016 Perturb-seq, K562 CRISPRi single perturbations (GEARS processing, Csendes et al. 2025)	0.797 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Adamson (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-adamson-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell F15, row 'Adamson / scGPT', column 'Pearson Delta DE'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Adamson et al. 2016 Perturb-seq, K562 CRISPRi single perturbations: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Adamson et al. 2016 Perturb-seq, K562 CRISPRi single perturbations (GEARS processing, Csendes et al. 2025)	0.795 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Adamson (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-adamson-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell H15, row 'Adamson / scGPT', column 'Pearson Delta DE without KOd gene'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Norman et al. 2019 Perturb-seq, K562 CRISPRa single and double perturbations: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Norman et al. 2019 Perturb-seq, K562 CRISPRa single and double perturbations (GEARS processing, Csendes et al. 2025)	0.988 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Norman (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-norman-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell C30, row 'Norman / scGPT', column 'Pearson'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Norman et al. 2019 Perturb-seq, K562 CRISPRa single and double perturbations: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Norman et al. 2019 Perturb-seq, K562 CRISPRa single and double perturbations (GEARS processing, Csendes et al. 2025)	0.935 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Norman (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-norman-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell D30, row 'Norman / scGPT', column 'Pearson DE'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Norman et al. 2019 Perturb-seq, K562 CRISPRa single and double perturbations: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Norman et al. 2019 Perturb-seq, K562 CRISPRa single and double perturbations (GEARS processing, Csendes et al. 2025)	0.554 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Norman (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-norman-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell E30, row 'Norman / scGPT', column 'Pearson Delta'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Norman et al. 2019 Perturb-seq, K562 CRISPRa single and double perturbations: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Norman et al. 2019 Perturb-seq, K562 CRISPRa single and double perturbations (GEARS processing, Csendes et al. 2025)	0.732 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Norman (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-norman-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell F30, row 'Norman / scGPT', column 'Pearson Delta DE'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Norman et al. 2019 Perturb-seq, K562 CRISPRa single and double perturbations: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Norman et al. 2019 Perturb-seq, K562 CRISPRa single and double perturbations (GEARS processing, Csendes et al. 2025)	0.7 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Norman (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-norman-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell H30, row 'Norman / scGPT', column 'Pearson Delta DE without KOd gene'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Replogle et al. 2022 genome-wide Perturb-seq subset, K562 CRISPRi: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Replogle et al. 2022 genome-wide Perturb-seq subset, K562 CRISPRi (GEARS processing, Csendes et al. 2025)	0.986 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Replogle_K562 (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-replogle-k562-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell C45, row 'Replogle_K562 / scGPT', column 'Pearson'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Replogle et al. 2022 genome-wide Perturb-seq subset, K562 CRISPRi: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Replogle et al. 2022 genome-wide Perturb-seq subset, K562 CRISPRi (GEARS processing, Csendes et al. 2025)	0.96 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Replogle_K562 (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-replogle-k562-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell D45, row 'Replogle_K562 / scGPT', column 'Pearson DE'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Replogle et al. 2022 genome-wide Perturb-seq subset, K562 CRISPRi: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Replogle et al. 2022 genome-wide Perturb-seq subset, K562 CRISPRi (GEARS processing, Csendes et al. 2025)	0.327 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Replogle_K562 (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-replogle-k562-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell E45, row 'Replogle_K562 / scGPT', column 'Pearson Delta'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Replogle et al. 2022 genome-wide Perturb-seq subset, K562 CRISPRi: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Replogle et al. 2022 genome-wide Perturb-seq subset, K562 CRISPRi (GEARS processing, Csendes et al. 2025)	0.481 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Replogle_K562 (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-replogle-k562-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell F45, row 'Replogle_K562 / scGPT', column 'Pearson Delta DE'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Replogle et al. 2022 genome-wide Perturb-seq subset, K562 CRISPRi: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Replogle et al. 2022 genome-wide Perturb-seq subset, K562 CRISPRi (GEARS processing, Csendes et al. 2025)	0.442 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Replogle_K562 (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-replogle-k562-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell H45, row 'Replogle_K562 / scGPT', column 'Pearson Delta DE without KOd gene'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Replogle et al. 2022 genome-wide Perturb-seq subset, RPE1 CRISPRi: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Replogle et al. 2022 genome-wide Perturb-seq subset, RPE1 CRISPRi (GEARS processing, Csendes et al. 2025)	0.974 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Replogle_RPE1 (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-replogle-rpe1-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell C60, row 'Replogle_RPE1 / scGPT', column 'Pearson'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Replogle et al. 2022 genome-wide Perturb-seq subset, RPE1 CRISPRi: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Replogle et al. 2022 genome-wide Perturb-seq subset, RPE1 CRISPRi (GEARS processing, Csendes et al. 2025)	0.924 pearson-correlation unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Replogle_RPE1 (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-replogle-rpe1-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell D60, row 'Replogle_RPE1 / scGPT', column 'Pearson DE'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Replogle et al. 2022 genome-wide Perturb-seq subset, RPE1 CRISPRi: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Replogle et al. 2022 genome-wide Perturb-seq subset, RPE1 CRISPRi (GEARS processing, Csendes et al. 2025)	0.596 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Replogle_RPE1 (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-replogle-rpe1-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell E60, row 'Replogle_RPE1 / scGPT', column 'Pearson Delta'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Replogle et al. 2022 genome-wide Perturb-seq subset, RPE1 CRISPRi: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Replogle et al. 2022 genome-wide Perturb-seq subset, RPE1 CRISPRi (GEARS processing, Csendes et al. 2025)	0.652 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Replogle_RPE1 (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-replogle-rpe1-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell F60, row 'Replogle_RPE1 / scGPT', column 'Pearson Delta DE'
Configuration: scGPT (Csendes et al. 2025)	Protocol: Replogle et al. 2022 genome-wide Perturb-seq subset, RPE1 CRISPRi: unseen-perturbation expression prediction (Csendes et al. 2025 Supplementary Table 2) Dataset: Replogle et al. 2022 genome-wide Perturb-seq subset, RPE1 CRISPRi (GEARS processing, Csendes et al. 2025)	0.629 pearson-delta unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Methods, coverage and source scGPT on Replogle_RPE1 (Csendes et al. 2025) perturbation-response-20261009-protocol-csendes2025-replogle-rpe1-pex Aggregation: Not reported Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Supplementary Table 2, cell H60, row 'Replogle_RPE1 / scGPT', column 'Pearson Delta DE without KOd gene'

Source checking is not independent reproduction. Release 2026-10-10-6e93f504adfc.

Use this model

How it works, versions and access

Strengths, limitations and unresolved questions

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 2026-10-10-6e93f504adfc
Property and statement	Original source and location	Review and provenance

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

All evidence (JSONL) (gzip)How to read the evidence

Sources and history

Release 2026-10-10-6e93f504adfc · Record review: source checked

2 source records and release history

Benchmarking foundation cell models for post-perturbation RNA-seq prediction · Original source · BMC Genomics 26:393, published 2025-04-23; PMC12016270 full-text XML
Csendes et al. 2025, Supplementary Material 4 (Supplementary Tables 1-3) · Original source · 12864_2025_11600_MOESM4_ESM.xlsx from the Europe PMC supplementary bundle (hash is of the workbook; the bundle zip is rebuilt per request)

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: perturbation-response-20261009-config-csendes2025-scgpt

areas: cells-tissues
contexts: research
method types: foundation_model
reported name: scGPT
source locator: Supplementary Table 2 column 'model'; Methods 'Foundation models' and 'Baseline models'
foundation model eligible: true
parameters: Pretrained scGPT fine-tuned per dataset for 15 epochs (best validation epoch), default hyperparameters (dimension 512, 12 layers, 8 heads), cell-gears v0.0.1 data processing
version: scGPT v0.2.1 fork, commit 7301b51
model identity note: Methods 'Foundation models' give 'a fork of scGPT v0.2.1 (commit 7301b51)'. The fork is not named, so the commit may not resolve in the upstream scGPT repository, and the pretrained checkpoint is given only as 'the pretrained models from the original publications'.

Related records

configuration of: scGPT
system: scGPT on Adamson (Csendes et al. 2025)
system: scGPT on Norman (Csendes et al. 2025)
system: scGPT on Replogle_K562 (Csendes et al. 2025)
system: scGPT on Replogle_RPE1 (Csendes et al. 2025)

Suggest a correction