| Configuration: DELFI features with Deep Learning (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.932 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceDELFI / Deep Learning on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'DELFI' / 'Deep Learning', column 'Performance (AUC)' |
|---|
| Configuration: DELFI features with Random Forest (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.921 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceDELFI / Random Forest on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'DELFI' / 'Random Forest', column 'Performance (AUC)' |
|---|
| Configuration: DELFI features with XGBoost (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.935 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceDELFI / XGBoost on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'DELFI' / 'XGBoost', column 'Performance (AUC)' |
|---|
| Configuration: FDI features with Deep Learning (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.851 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceFDI / Deep Learning on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'FDI' / 'Deep Learning', column 'Performance (AUC)' |
|---|
| Configuration: FDI features with Random Forest (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.847 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceFDI / Random Forest on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'FDI' / 'Random Forest', column 'Performance (AUC)' |
|---|
| Configuration: FDI features with XGBoost (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.844 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceFDI / XGBoost on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'FDI' / 'XGBoost', column 'Performance (AUC)' |
|---|
| Configuration: FRAGMA features with Deep Learning (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.853 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceFRAGMA / Deep Learning on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'FRAGMA' / 'Deep Learning', column 'Performance (AUC)' |
|---|
| Configuration: FRAGMA features with Random Forest (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.838 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceFRAGMA / Random Forest on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'FRAGMA' / 'Random Forest', column 'Performance (AUC)' |
|---|
| Configuration: FRAGMA features with XGBoost (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.849 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceFRAGMA / XGBoost on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'FRAGMA' / 'XGBoost', column 'Performance (AUC)' |
|---|
| Configuration: Fragmentia-AI trained on ULP-WGS (separate instance) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.956 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · Source checkedMethods, coverage and sourcecfDNA sequence / Fragmentia-AI on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'cfDNA sequence' / 'Fragmentia-AI', column 'Performance (AUC)' |
|---|
| Configuration: Griffin features with Deep Learning (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.885 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceGriffin / Deep Learning on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'Griffin' / 'Deep Learning', column 'Performance (AUC)' |
|---|
| Configuration: Griffin features with Random Forest (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.881 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceGriffin / Random Forest on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'Griffin' / 'Random Forest', column 'Performance (AUC)' |
|---|
| Configuration: Griffin features with XGBoost (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.886 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceGriffin / XGBoost on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'Griffin' / 'XGBoost', column 'Performance (AUC)' |
|---|
| Configuration: Nucleosome features with Deep Learning (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.81 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceNucleosome / Deep Learning on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'Nucleosome' / 'Deep Learning', column 'Performance (AUC)' |
|---|
| Configuration: Nucleosome features with Random Forest (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.793 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceNucleosome / Random Forest on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'Nucleosome' / 'Random Forest', column 'Performance (AUC)' |
|---|
| Configuration: Nucleosome features with XGBoost (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.801 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceNucleosome / XGBoost on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'Nucleosome' / 'XGBoost', column 'Performance (AUC)' |
|---|
| Configuration: Subnucleosome features with Deep Learning (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.636 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceSubnucleosome / Deep Learning on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'Subnucleosome' / 'Deep Learning', column 'Performance (AUC)' |
|---|
| Configuration: Subnucleosome features with Random Forest (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.674 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceSubnucleosome / Random Forest on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'Subnucleosome' / 'Random Forest', column 'Performance (AUC)' |
|---|
| Configuration: Subnucleosome features with XGBoost (Fragmentia-AI paper re-implementation) | Protocol: Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers Dataset: Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer) | 0.676 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible | Independent external evaluation · Source checkedMethods, coverage and sourceSubnucleosome / XGBoost on the ULP-WGS held-out validation set (Fragmentia-AI paper) ctdnalit-20261011-fragmentia2026-protocol-ulp-wgs-heldout-auc Aggregation: Not reported Toward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'Subnucleosome' / 'XGBoost', column 'Performance (AUC)' |
|---|