Result

Rank (rank among the eight representations by mean biological-conservation and batch-mixing scores on this dataset (1 = best); full embedding dimensions) of Harmony (Kedzierska et al. 2025) on PBMC (12k) (Kedzierska et al. 2025)

1 Rank (rank among the eight representations by mean biological-conservation and batch-mixing scores on this dataset (1 = best); full embedding dimensions)

Methods

Tested configuration: Harmony (Kedzierska et al. 2025)
Protocol: Kedzierska et al. 2025 cell embedding scoring, full embedding dimensions
Dataset subset: PBMC (12k) (Kedzierska et al. 2025)
Procedure: Kedzierska et al. 2025 cell embedding scoring, full embedding dimensions
Evaluation: Harmony cell embedding on PBMC (12k) (Kedzierska et al. 2025)
Coverage: Not reported scored / Not reported eligible
Uncertainty: Not reported by the source
Evidence: Independent external evaluation · source checkedKedzierska et al. 2025, Additional file 1 (Supplementary Tables S1-S7, xlsx) · Additional file 1, sheet 'TableS4' cell D4, row 'Harmony', column 'PBMC (12k)'

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: scib 1.0.4. ASW of cell-type labels rescaled to [0, 1]; NMI and ARI on Louvain clusters at the resolution with the highest NMI out of 20; batch integration = 1 minus absolute batch silhouette per cell type, averaged; PCR = change in variance explained by batch between the original data and the embedding (Methods, pages 9-10).

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

Claims, original sources and review scope · Release 2026-10-23-5347695500ed
Property and statement	Original source and location	Review and provenance
Reported result 1 Individual claims	Kedzierska et al. 2025, Additional file 1 (Supplementary Tables S1-S7, xlsx) Original source ↗ Additional file 1, sheet 'TableS4' cell D4, row 'Harmony', column 'PBMC (12k)' Version: Additional file 1 (13059_2025_3574_MOESM1_ESM.xlsx) of the version of record Retrieved: 2026-10-14T12:01:36Z	source checked ["source-hash-verification","deterministic-table-parse","independent-cell-check"] · 2026-10-14 independent paper Audit details Extraction: Read sheet 'TableS4' of the pinned workbook with a stdlib OOXML reader; caption, header row A2:G2 and model labels A3:A10 asserted; printed_value is the Excel General display of the stored number. Independent review 2026-10-14: matched. The single-value columns are whole-dataset scores: medians of these columns across the five main datasets reproduce 54 of the 56 TableS3 ranks (the bracketed columns reproduce 39). Field: `attributes.printed_value` Source artifact SHA-256: `7adee1211206c4071981a402617c293048baeeeaf415b68ecd13383678003b5e` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact Extraction artifact SHA-256: `7adee1211206c4071981a402617c293048baeeeaf415b68ecd13383678003b5e` Extraction artifact

How to read the evidence

Sources and history

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

2 source records and release history

Kedzierska et al. 2025, Additional file 1 (Supplementary Tables S1-S7, xlsx) · Original source · Additional file 1 (13059_2025_3574_MOESM1_ESM.xlsx) of the version of record
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

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: gfm-guides-20261014-kedzierska2025-result-emb-pbmc-12k-harmony-rank-average

metric: rank
metric direction: lower
unit: unitless
printed value: 1
numeric value: 1
metric qualifier: rank among the eight representations by mean biological-conservation and batch-mixing scores on this dataset (1 = best); full embedding dimensions
source locator: Additional file 1, sheet 'TableS4' cell D4, row 'Harmony', column 'PBMC (12k)'
review: method: source-hash-verification; deterministic-table-parse; independent-cell-check; method note: Workbook re-downloaded from the publisher (SHA-256 matches). An independent openpyxl reader compared every one of the 600 workbook results with the cell its locator names: value, Excel General display, metric block, dataset row, model column, protocol (full or UMAP) and evaluation links; every expected cell was found exactly once. Six cells also checked by hand.; reviewer: claude; reviewer note: Claude (Opus 5.5) research agent, independent of the extractor; no human review claimed; date: 2026-10-14; artifact sha256: 7adee1211206c4071981a402617c293048baeeeaf415b68ecd13383678003b5e; retrieval url: https://static-content.springer.com/esm/art%3A10.1186%2Fs13059-025-03574-x/MediaObjects/13059_2025_3574_MOESM1_ESM.xlsx; note: Extraction: Read sheet 'TableS4' of the pinned workbook with a stdlib OOXML reader; caption, header row A2:G2 and model labels A3:A10 asserted; printed_value is the Excel General display of the stored number. Independent review 2026-10-14: matched. The single-value columns are whole-dataset scores: medians of these columns across the five main datasets reproduce 54 of the 56 TableS3 ranks (the bracketed columns reproduce 39).
missing metadata: uncertainty: reason: unreported
unit detail: ordinal position among eight representations; ties share a rank
raw xml value: 1.0
workbook number format: General

Related records

evaluation: Harmony cell embedding on PBMC (12k) (Kedzierska et al. 2025)

Suggest a correction