Protocol

SCMBench paired RNA-ATAC integration accuracy

Evidence concern: excluded from comparisons

Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H

'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows

In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20

Integrate paired scRNA-seq and scATAC-seq; score the joint embedding by the mean of mean average precision, normalised mutual information, average silhouette width and ARI of cell type labels (integration accuracy overall score).

128 evaluations · 128 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.

128 recorded evaluations, 128 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

128 evaluations · 128 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: bindSC (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.66 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source bindSC on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B5 (row 'bindSC', column '10x-Multiome-Pbmc10k')
Configuration: Cobolt (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.631 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source Cobolt on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B7 (row 'Cobolt', column '10x-Multiome-Pbmc10k')
Configuration: Deepmaps (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.567 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source Deepmaps on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B16 (row 'Deepmaps', column '10x-Multiome-Pbmc10k')
Configuration: FM-scVI (SCMBench adapter)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.674 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source FM-scVI on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B23 (row 'FM-scVI', column '10x-Multiome-Pbmc10k')
Configuration: GLUE (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.676 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source GLUE on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B2 (row 'GLUE', column '10x-Multiome-Pbmc10k')
Configuration: Harmony (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.574 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source Harmony on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B22 (row 'Harmony', column '10x-Multiome-Pbmc10k')
Configuration: iNMF (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.203 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source iNMF on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B21 (row 'iNMF', column '10x-Multiome-Pbmc10k')
Configuration: LIGER (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.509 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source LIGER on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B11 (row 'LIGER', column '10x-Multiome-Pbmc10k')
Configuration: MMD-MA (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.331 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source MMD-MA on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B17 (row 'MMD-MA', column '10x-Multiome-Pbmc10k')
Configuration: MOFA (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.535 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source MOFA on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B10 (row 'MOFA', column '10x-Multiome-Pbmc10k')
Configuration: Pamona (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.472 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source Pamona on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B19 (row 'Pamona', column '10x-Multiome-Pbmc10k')
Configuration: PCA (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.48 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source PCA on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B14 (row 'PCA', column '10x-Multiome-Pbmc10k')
Configuration: scGPT (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.591 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source scGPT on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B20 (row 'scGPT', column '10x-Multiome-Pbmc10k')
Configuration: scJoint (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.767 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source scJoint on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B4 (row 'scJoint', column '10x-Multiome-Pbmc10k')
Configuration: scMDC (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.739 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source scMDC on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B3 (row 'scMDC', column '10x-Multiome-Pbmc10k')
Configuration: scMoMaT (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.628 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source scMoMaT on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B9 (row 'scMoMaT', column '10x-Multiome-Pbmc10k')
Configuration: scVI (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.551 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source scVI on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B15 (row 'scVI', column '10x-Multiome-Pbmc10k')
Configuration: Seurat4 (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.795 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source Seurat4 on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B6 (row 'Seurat4', column '10x-Multiome-Pbmc10k')
Configuration: Seurat5 (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.552 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source Seurat5 on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B8 (row 'Seurat5', column '10x-Multiome-Pbmc10k')
Configuration: TotalVI (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.506 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source TotalVI on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B13 (row 'TotalVI', column '10x-Multiome-Pbmc10k')
Configuration: UCE (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.533 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source UCE on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B18 (row 'UCE', column '10x-Multiome-Pbmc10k')
Configuration: UnionCom (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: 10x-Multiome-Pbmc10k (SCMBench)	0.453 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source UnionCom on SCMBench paired integration, 10x-Multiome-Pbmc10k lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell B12 (row 'UnionCom', column '10x-Multiome-Pbmc10k')
Configuration: bindSC (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: Chen-2019 (SCMBench)	0.352 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source bindSC on SCMBench paired integration, Chen-2019 lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell C5 (row 'bindSC', column 'Chen-2019')
Configuration: Cobolt (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: Chen-2019 (SCMBench)	0.546 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source Cobolt on SCMBench paired integration, Chen-2019 lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell C7 (row 'Cobolt', column 'Chen-2019')
Configuration: Deepmaps (SCMBench)	Protocol: SCMBench paired RNA-ATAC integration accuracy Dataset subset: Chen-2019 (SCMBench)	0.484 Scmbench integration accuracy (mean of MAP, NMI, ASW and ARI) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Independent external evaluation · Source checked Evidence concern: excluded from comparisons Sheet 'unpaired_inte_overall' has no header for column H. Each H cell is the formula =AVERAGE(Bn:Gn), the mean of the dataset cells with values. Rows with a blank Yao cell (bindSC, scMoMaT, scGPT_zero, Pamona, scFoundation, Geneformer) average five datasets and the other rows six, so column H values are not comparable across all rows.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'unpaired_inte_overall', row 1 and column H 'FM-scVI' is a different foundation model per dataset ('the best-performing foundation model combined with scVI, which varies across datasets', Fig. 2 legend), so its rows do not describe one fixed configuration.SCMBench Supplementary Data 2 (integration results) · Article Fig. 2 legend; Supplementary Data 2 'FM-scVI' rows In sheet 'paired_inte_overall' the scGPT row is scored on 10x-Multiome-Pbmc10k only; its Overall cell G20 (=AVERAGE(B20:F20)) is that single value, not a five-dataset mean, so it is not comparable with the other Overall values.SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell G20 Methods, coverage and source Deepmaps on SCMBench paired integration, Chen-2019 lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy Aggregation: Not reported SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration; SCMBench Supplementary Data 2 (integration results) · Supplementary Data 2, sheet 'paired_inte_overall', cell C16 (row 'Deepmaps', column 'Chen-2019')

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.

Explore all linked results

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: 6
External evaluations: 122

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

Null control

Proposed control: requires review

Unintegrated representation using protocol preprocessing

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

Protocol-specific conventional integration method

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

2 source records and release history

SCMBench: benchmarking domain-specific and foundation models for single-cell multi-omics data integration · Original source · Published article, Nature Communications 2026 (PMC13342640.1)
SCMBench Supplementary Data 2 (integration results) · Original source · Supplementary Data 2 (MOESM4) of the published article

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: lit-cells-networks-20261011-scmbench-protocol-paired-integration-accuracy

areas: single-cell
contexts: research
protocol: Paired multi-omics integration; per-dataset overall score is the mean of MAP, NMI, ASW and ARI ('The overall score represents the mean of these four integration accuracy metrics', Results).
version: Nature Communications 2026
metric: scmbench-integration-accuracy
metric direction: higher
unit: unitless
limitations: Single run per method and dataset; no spread printed.; Foundation models receive ATAC as gene activity scores (MAESTRO).
source locator: Results 'Benchmarking data integration on both paired and unpaired datasets'; Fig. 2e-f; Supplementary Data 2

Related records

part of: SCMBench
assessment: bindSC on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: Cobolt on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: Deepmaps on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: FM-scVI on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: GLUE on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: Harmony on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: iNMF on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: LIGER on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: MMD-MA on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: MOFA on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: Pamona on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: PCA on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: scGPT on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: scJoint on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: scMDC on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: scMoMaT on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: scVI on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: Seurat4 on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: Seurat5 on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: TotalVI on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: UCE on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: UnionCom on SCMBench paired integration, 10x-Multiome-Pbmc10k
assessment: bindSC on SCMBench paired integration, Chen-2019
assessment: Cobolt on SCMBench paired integration, Chen-2019
assessment: Deepmaps on SCMBench paired integration, Chen-2019
assessment: FM-scVI on SCMBench paired integration, Chen-2019
assessment: GLUE on SCMBench paired integration, Chen-2019
assessment: Harmony on SCMBench paired integration, Chen-2019
assessment: iNMF on SCMBench paired integration, Chen-2019
assessment: LIGER on SCMBench paired integration, Chen-2019
assessment: MMD-MA on SCMBench paired integration, Chen-2019
assessment: MOFA on SCMBench paired integration, Chen-2019
assessment: Pamona on SCMBench paired integration, Chen-2019
assessment: PCA on SCMBench paired integration, Chen-2019
assessment: scJoint on SCMBench paired integration, Chen-2019
assessment: scMDC on SCMBench paired integration, Chen-2019
assessment: scMoMaT on SCMBench paired integration, Chen-2019
assessment: scVI on SCMBench paired integration, Chen-2019
assessment: Seurat4 on SCMBench paired integration, Chen-2019
assessment: Seurat5 on SCMBench paired integration, Chen-2019
assessment: TotalVI on SCMBench paired integration, Chen-2019
assessment: UCE on SCMBench paired integration, Chen-2019
assessment: UnionCom on SCMBench paired integration, Chen-2019
assessment: bindSC on SCMBench paired integration, Ma-2020-batch-53
assessment: Cobolt on SCMBench paired integration, Ma-2020-batch-53
assessment: Deepmaps on SCMBench paired integration, Ma-2020-batch-53
assessment: FM-scVI on SCMBench paired integration, Ma-2020-batch-53
assessment: GLUE on SCMBench paired integration, Ma-2020-batch-53
assessment: Harmony on SCMBench paired integration, Ma-2020-batch-53
assessment: iNMF on SCMBench paired integration, Ma-2020-batch-53
assessment: LIGER on SCMBench paired integration, Ma-2020-batch-53
assessment: MMD-MA on SCMBench paired integration, Ma-2020-batch-53
assessment: MOFA on SCMBench paired integration, Ma-2020-batch-53
assessment: Pamona on SCMBench paired integration, Ma-2020-batch-53
assessment: PCA on SCMBench paired integration, Ma-2020-batch-53
assessment: scJoint on SCMBench paired integration, Ma-2020-batch-53
assessment: scMDC on SCMBench paired integration, Ma-2020-batch-53
assessment: scMoMaT on SCMBench paired integration, Ma-2020-batch-53
assessment: scVI on SCMBench paired integration, Ma-2020-batch-53
assessment: Seurat4 on SCMBench paired integration, Ma-2020-batch-53
assessment: Seurat5 on SCMBench paired integration, Ma-2020-batch-53
assessment: TotalVI on SCMBench paired integration, Ma-2020-batch-53
assessment: UCE on SCMBench paired integration, Ma-2020-batch-53
assessment: UnionCom on SCMBench paired integration, Ma-2020-batch-53
assessment: bindSC on SCMBench paired integration, Ma-2020-batch-54
assessment: Cobolt on SCMBench paired integration, Ma-2020-batch-54
assessment: Deepmaps on SCMBench paired integration, Ma-2020-batch-54
assessment: FM-scVI on SCMBench paired integration, Ma-2020-batch-54
assessment: GLUE on SCMBench paired integration, Ma-2020-batch-54
assessment: Harmony on SCMBench paired integration, Ma-2020-batch-54
assessment: iNMF on SCMBench paired integration, Ma-2020-batch-54
assessment: LIGER on SCMBench paired integration, Ma-2020-batch-54
assessment: MMD-MA on SCMBench paired integration, Ma-2020-batch-54
assessment: MOFA on SCMBench paired integration, Ma-2020-batch-54
assessment: Pamona on SCMBench paired integration, Ma-2020-batch-54
assessment: PCA on SCMBench paired integration, Ma-2020-batch-54
assessment: scJoint on SCMBench paired integration, Ma-2020-batch-54
assessment: scMDC on SCMBench paired integration, Ma-2020-batch-54
assessment: scMoMaT on SCMBench paired integration, Ma-2020-batch-54
assessment: scVI on SCMBench paired integration, Ma-2020-batch-54
assessment: Seurat4 on SCMBench paired integration, Ma-2020-batch-54
assessment: Seurat5 on SCMBench paired integration, Ma-2020-batch-54
assessment: TotalVI on SCMBench paired integration, Ma-2020-batch-54
assessment: UCE on SCMBench paired integration, Ma-2020-batch-54
assessment: UnionCom on SCMBench paired integration, Ma-2020-batch-54
assessment: bindSC on SCMBench paired integration, Ma-2020-batch-55
assessment: Cobolt on SCMBench paired integration, Ma-2020-batch-55
assessment: Deepmaps on SCMBench paired integration, Ma-2020-batch-55
assessment: FM-scVI on SCMBench paired integration, Ma-2020-batch-55
assessment: GLUE on SCMBench paired integration, Ma-2020-batch-55
assessment: Harmony on SCMBench paired integration, Ma-2020-batch-55
assessment: iNMF on SCMBench paired integration, Ma-2020-batch-55
assessment: LIGER on SCMBench paired integration, Ma-2020-batch-55
assessment: MMD-MA on SCMBench paired integration, Ma-2020-batch-55
assessment: MOFA on SCMBench paired integration, Ma-2020-batch-55
assessment: Pamona on SCMBench paired integration, Ma-2020-batch-55
assessment: PCA on SCMBench paired integration, Ma-2020-batch-55
assessment: scJoint on SCMBench paired integration, Ma-2020-batch-55
assessment: scMDC on SCMBench paired integration, Ma-2020-batch-55
assessment: scMoMaT on SCMBench paired integration, Ma-2020-batch-55
assessment: scVI on SCMBench paired integration, Ma-2020-batch-55
assessment: Seurat4 on SCMBench paired integration, Ma-2020-batch-55
assessment: Seurat5 on SCMBench paired integration, Ma-2020-batch-55
assessment: TotalVI on SCMBench paired integration, Ma-2020-batch-55
assessment: UCE on SCMBench paired integration, Ma-2020-batch-55
assessment: UnionCom on SCMBench paired integration, Ma-2020-batch-55
assessment: bindSC on SCMBench paired integration, paired-overall
assessment: Cobolt on SCMBench paired integration, paired-overall
assessment: Deepmaps on SCMBench paired integration, paired-overall
assessment: FM-scVI on SCMBench paired integration, paired-overall
assessment: GLUE on SCMBench paired integration, paired-overall
assessment: Harmony on SCMBench paired integration, paired-overall
assessment: iNMF on SCMBench paired integration, paired-overall
assessment: LIGER on SCMBench paired integration, paired-overall
assessment: MMD-MA on SCMBench paired integration, paired-overall
assessment: MOFA on SCMBench paired integration, paired-overall
assessment: Pamona on SCMBench paired integration, paired-overall
assessment: PCA on SCMBench paired integration, paired-overall
assessment: scGPT on SCMBench paired integration, paired-overall
assessment: scJoint on SCMBench paired integration, paired-overall
assessment: scMDC on SCMBench paired integration, paired-overall
assessment: scMoMaT on SCMBench paired integration, paired-overall
assessment: scVI on SCMBench paired integration, paired-overall
assessment: Seurat4 on SCMBench paired integration, paired-overall
assessment: Seurat5 on SCMBench paired integration, paired-overall
assessment: TotalVI on SCMBench paired integration, paired-overall
assessment: UCE on SCMBench paired integration, paired-overall
assessment: UnionCom on SCMBench paired integration, paired-overall

Suggest a correction