Pearson (Geneformer input versus predicted gene ranking, all genes unmasked; best value across datasets) of Geneformer (Kedzierska et al. 2025) on Evaluated datasets, aggregated (Kedzierska et al. 2025)
0.95 Pearson (Geneformer input versus predicted gene ranking, all genes unmasked; best value across datasets)
Methods
- Tested configuration
- Geneformer (Kedzierska et al. 2025)
- Protocol
- Kedzierska et al. 2025 Geneformer pretraining-objective reconstruction
- Dataset subset
- Evaluated datasets, aggregated (Kedzierska et al. 2025)
- Procedure
- Kedzierska et al. 2025 Geneformer pretraining-objective reconstruction
- Evaluation
- Geneformer pretraining reconstruction on evaluated datasets (Kedzierska et al. 2025)
- Coverage
- Not reported scored / Not reported eligible
- Uncertainty
- Not reported by the source
- Evidence
- Independent external evaluation · source checkedZero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) · Page 5, 'Results and discussion', paragraph beginning 'Notably, scGPT without embeddings', sentence 2 ('median and best correlations of 0.56 and 0.95, respectively, across datasets')
A source-checked result verifies the numerical transcription, not every model or protocol detail. Evaluation metadata: source checked. Source checked does not mean independently reproduced.
Reproduction
- Split
- No split; each whole dataset embedded and scored
- Adaptation
- scGPT and Geneformer zero-shot from pretrained weights (scGPT random: no pretraining); scVI and Harmony fitted on each dataset with batch labels; HVG computed from each dataset
- Scoring implementation
- Pearson correlation of true and predicted scaled gene rankings (Methods page 10)
No execution recipe has been verified for this exact configuration and evaluation. A benchmark's general instructions may use different inputs, splits or model settings.
Reproducing this published result requires matching its model configuration, data, split and scorer. Source checking or a successful smoke test does not establish score reproduction.
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.
1 evidence row matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Reported result 0.95 Individual claims | Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) Page 5, 'Results and discussion', paragraph beginning 'Notably, scGPT without embeddings', sentence 2 ('median and best correlations of 0.56 and 0.95, respectively, across datasets') Version: Version of record, Genome Biology 26:101, published 2025-04-18 | source checked ["source-hash-verification","pdf-text-parse","independent-cell-check"] · 2026-10-14 independent paper Audit detailsExtraction: Hand transcription from the article text; the script asserts the sentence in the pdftotext -layout output of page 5 of the pinned PDF. Independent review 2026-10-14: matched. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record Extraction artifact SHA-256: |
Sources and history
Release 2026-10-23-5347695500ed · Record review: source checked
1 source record and release history
- Zero-shot evaluation reveals limitations of single-cell foundation models (Kedzierska et al. 2025, Genome Biology 26:101) · Original source · Version of record, Genome Biology 26:101, published 2025-04-18
Technical metadata and extraction receipts
Stable ID: gfm-guides-20261014-kedzierska2025-result-recon-evaluated-datasets-geneformer-pearson-best
- metric
- pearson-correlation
- metric direction
- higher
- unit
- unitless
- printed value
- 0.95
- numeric value
- 0.95
- metric qualifier
- Geneformer input versus predicted gene ranking, all genes unmasked; best value across datasets
- source locator
- Page 5, 'Results and discussion', paragraph beginning 'Notably, scGPT without embeddings', sentence 2 ('median and best correlations of 0.56 and 0.95, respectively, across datasets')
- review
- method: source-hash-verification; pdf-text-parse; independent-cell-check; method note: Sentence re-read in the pdftotext layer of the archived article PDF (SHA-256 matches the pinned hash).; reviewer: claude; reviewer note: Claude (Opus 5.5) research agent, independent of the extractor; no human review claimed; date: 2026-10-14; artifact sha256: 8efcc5fcb2e35b9a04b1ee9e03581e20fffeb5521a4a4c748a6f9b34f7d3c47f; retrieval url: https://genomebiology.biomedcentral.com/counter/pdf/10.1186/s13059-025-03574-x.pdf; note: Extraction: Hand transcription from the article text; the script asserts the sentence in the pdftotext -layout output of page 5 of the pinned PDF. Independent review 2026-10-14: matched.
- missing metadata
- uncertainty: reason: unreported