rewirebio.iobenchmarks
Contested result

MutationAssessor and PrimateAI: Comprehensive assessment of computational algorithms in predicting cancer driver mutations and Livesey and Marsh 2025, Genome Biology 26:104 (Europe PMC full text XML) disagree

  • Low severity
  • Not yet reviewed
  • Two-study conflict
  • unknown

Comprehensive assessment of computational algorithms in predicting cancer driver mutations ranks PrimateAI above MutationAssessor in 19 of 20 comparisons; Livesey and Marsh 2025, Genome Biology 26:104 (Europe PMC full text XML) ranks MutationAssessor above PrimateAI in 7 of 7 comparisons. No other study gives a verdict, so neither is the outlier.

An automatic lead, not a finding of error. The results below are shown as published, everywhere on the site, until a reviewer decides otherwise.

Which studies agree

5 papers rank MutationAssessor above PrimateAI in 15 of their 40 comparisons and PrimateAI higher in 24; 1 is a tie. Of those papers, Comprehensive assessment… ranks PrimateAI higher on its own (19 of 20).

Papers grouped by which system they rank higher: MutationAssessor or PrimateAIRank MutationAssessor higher: Livesey and Marsh 2025, Clinical likelihood ratios and balanced…, ProteinGym. Rank PrimateAI higher: Comprehensive assessment…. Mixed: An openly licensed benchmark….Rank MutationAssessor higherMixedRank PrimateAI higherLivesey and Marsh 2025Livesey and Marsh 2025, Genome Biology 26:104 (Europe PMC full text XML)Clinical likelihoodratios and balanced…Clinical likelihood ratios and balanced accuracy for 44 in silico tools against multiple large-scale functional assays of cancer susceptibility genesProteinGymProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and DesignAn openly licensedbenchmark…An openly licensed benchmark and per-gene calibration map for missense pathogenicity predictors on activating cancer driversComprehensive assessment…Comprehensive assessment of computational algorithms in predicting cancer driver mutations
  • Rank MutationAssessor higher (3): Livesey and Marsh 2025, Clinical likelihood ratios and balanced…, ProteinGym
  • Rank PrimateAI higher (1): Comprehensive assessment…
  • Mixed (1): An openly licensed benchmark…

Each paper sits where at least two thirds of its comparisons point; Mixed means neither model reaches two thirds. The larger, outlined dot is the paper that disagrees.

Comparisons in which each paper scores MutationAssessor or PrimateAI higherComprehensive assessment…: MutationAssessor higher in 1, PrimateAI higher in 19, 0 ties. Livesey and Marsh 2025: MutationAssessor higher in 7, PrimateAI higher in 0, 0 ties. Clinical likelihood ratios and balanced…: MutationAssessor higher in 4, PrimateAI higher in 2, 0 ties. An openly licensed benchmark…: MutationAssessor higher in 2, PrimateAI higher in 3, 1 tie. ProteinGym: MutationAssessor higher in 1, PrimateAI higher in 0, 0 ties.Comprehensive assessment…Comprehensive assessment of computational algorithms in predicting cancer driver mutations1 MutationAssessor higher19 PrimateAI higher19Livesey and Marsh 2025Livesey and Marsh 2025, Genome Biology 26:104 (Europe PMC full text XML)7 MutationAssessor higher7Clinical likelihood ratios and balanced…Clinical likelihood ratios and balanced accuracy for 44 in silico tools against multiple large-scale functional assays of cancer susceptibility genes4 MutationAssessor higher42 PrimateAI higher2An openly licensed benchmark…An openly licensed benchmark and per-gene calibration map for missense pathogenicity predictors on activating cancer drivers2 MutationAssessor higher21 tie13 PrimateAI higher3ProteinGymProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design1 MutationAssessor higher1
MutationAssessor higherTiePrimateAI higherNumbers are counts of comparisons; the outlined bar is the paper that disagrees.
Comparisons in Comprehensive assessment… (20)
ProtocolDatasetMetricMutationAssessorPrimateAIHigher
OncoKB benchmark (positive set not stated), default categorical calls (Chen et al. 2020 Additional file 10)OncoKB-annotated somatic missense mutations (oncogenic, likely oncogenic and likely neutral)Accuracy (OncoKB benchmark (positive set not stated), default categorical calls), fraction0.580.66PrimateAI
OncoKB benchmark (positive set not stated), default categorical calls (Chen et al. 2020 Additional file 10)OncoKB-annotated somatic missense mutations (oncogenic, likely oncogenic and likely neutral)Negative predictive value (OncoKB benchmark (positive set not stated), default categorical calls), fraction0.590.71PrimateAI
OncoKB benchmark (positive set not stated), default categorical calls (Chen et al. 2020 Additional file 10)OncoKB-annotated somatic missense mutations (oncogenic, likely oncogenic and likely neutral)Precision (OncoKB benchmark (positive set not stated), default categorical calls), fraction0.570.63PrimateAI
OncoKB benchmark (positive set not stated), default categorical calls (Chen et al. 2020 Additional file 10)OncoKB-annotated somatic missense mutations (oncogenic, likely oncogenic and likely neutral)Recall (OncoKB benchmark (positive set not stated), default categorical calls), fraction0.660.77PrimateAI
OncoKB benchmark (positive set not stated), default categorical calls (Chen et al. 2020 Additional file 10)OncoKB-annotated somatic missense mutations (oncogenic, likely oncogenic and likely neutral)Specificity (OncoKB benchmark (positive set not stated), default categorical calls), fraction0.50.55PrimateAI
OncoKB benchmark (positive set not stated), median-score threshold (Chen et al. 2020 Additional file 9)OncoKB-annotated somatic missense mutations (oncogenic, likely oncogenic and likely neutral)Accuracy (OncoKB benchmark (positive set not stated), median-score threshold), fraction0.590.66PrimateAI
OncoKB benchmark (positive set not stated), median-score threshold (Chen et al. 2020 Additional file 9)OncoKB-annotated somatic missense mutations (oncogenic, likely oncogenic and likely neutral)Negative predictive value (OncoKB benchmark (positive set not stated), median-score threshold), fraction0.590.66PrimateAI
OncoKB benchmark (positive set not stated), median-score threshold (Chen et al. 2020 Additional file 9)OncoKB-annotated somatic missense mutations (oncogenic, likely oncogenic and likely neutral)Precision (OncoKB benchmark (positive set not stated), median-score threshold), fraction0.590.66PrimateAI
OncoKB benchmark (positive set not stated), median-score threshold (Chen et al. 2020 Additional file 9)OncoKB-annotated somatic missense mutations (oncogenic, likely oncogenic and likely neutral)Recall (OncoKB benchmark (positive set not stated), median-score threshold), fraction0.610.66PrimateAI
OncoKB benchmark (positive set not stated), median-score threshold (Chen et al. 2020 Additional file 9)OncoKB-annotated somatic missense mutations (oncogenic, likely oncogenic and likely neutral)Specificity (OncoKB benchmark (positive set not stated), median-score threshold), fraction0.570.66PrimateAI
Cell viability drivers versus neutral, default categorical calls (Chen et al. 2020 Additional file 22)Missense mutations with Ba/F3 and MCF10A cell viability calls (published and new)Accuracy (cell viability drivers versus neutral, default categorical calls), fraction0.60.65PrimateAI
Cell viability drivers versus neutral, default categorical calls (Chen et al. 2020 Additional file 22)Missense mutations with Ba/F3 and MCF10A cell viability calls (published and new)Negative predictive value (cell viability drivers versus neutral, default categorical calls), fraction0.590.7PrimateAI
Cell viability drivers versus neutral, default categorical calls (Chen et al. 2020 Additional file 22)Missense mutations with Ba/F3 and MCF10A cell viability calls (published and new)Precision (cell viability drivers versus neutral, default categorical calls), fraction0.60.63PrimateAI
Cell viability drivers versus neutral, default categorical calls (Chen et al. 2020 Additional file 22)Missense mutations with Ba/F3 and MCF10A cell viability calls (published and new)Recall (cell viability drivers versus neutral, default categorical calls), fraction0.60.79PrimateAI
Cell viability drivers versus neutral, default categorical calls (Chen et al. 2020 Additional file 22)Missense mutations with Ba/F3 and MCF10A cell viability calls (published and new)Specificity (cell viability drivers versus neutral, default categorical calls), fraction0.60.52MutationAssessor
Cell viability drivers versus neutral, median-score threshold (Chen et al. 2020 Additional file 21)Missense mutations with Ba/F3 and MCF10A cell viability calls (published and new)Accuracy (cell viability drivers versus neutral, median-score threshold), fraction0.60.7PrimateAI
Cell viability drivers versus neutral, median-score threshold (Chen et al. 2020 Additional file 21)Missense mutations with Ba/F3 and MCF10A cell viability calls (published and new)Negative predictive value (cell viability drivers versus neutral, median-score threshold), fraction0.60.7PrimateAI
Cell viability drivers versus neutral, median-score threshold (Chen et al. 2020 Additional file 21)Missense mutations with Ba/F3 and MCF10A cell viability calls (published and new)Precision (cell viability drivers versus neutral, median-score threshold), fraction0.60.7PrimateAI
Cell viability drivers versus neutral, median-score threshold (Chen et al. 2020 Additional file 21)Missense mutations with Ba/F3 and MCF10A cell viability calls (published and new)Recall (cell viability drivers versus neutral, median-score threshold), fraction0.60.7PrimateAI
Cell viability drivers versus neutral, median-score threshold (Chen et al. 2020 Additional file 21)Missense mutations with Ba/F3 and MCF10A cell viability calls (published and new)Specificity (cell viability drivers versus neutral, median-score threshold), fraction0.60.7PrimateAI
Comparisons in Livesey and Marsh 2025 (7)
ProtocolDatasetMetricMutationAssessorPrimateAIHigher
Livesey and Marsh 2025 clinical benchmark: pairwise AUROC win rate, ClinVar pathogenic versus gnomAD missenseLivesey and Marsh 2025 clinical set: ClinVar pathogenic versus gnomAD v4.1 missense (819 proteins)Pairwise win rate (average pairwise win rate against all other VEPs, per-protein AUROC on shared variants, ClinVar P/LP versus gnomAD v4.1 missense, proteins with at least 10 of each), percent51.431.6MutationAssessor
Livesey and Marsh 2025 DMS benchmark: pairwise |Spearman| win rate, all 36 DMS setsLivesey and Marsh 2025 human DMS benchmark set (36 proteins)Pairwise win rate (average pairwise win rate against all other VEPs, |Spearman| with DMS on shared variants, 36 human DMS sets), percent58.819.9MutationAssessor
Livesey and Marsh 2025 DMS benchmark: pairwise |Spearman| win rate, all 36 DMS setsLivesey and Marsh 2025 human DMS benchmark set (36 proteins)Pairwise win rate (average pairwise win rate against the five top-ranked VEPs (column top_5_winrate; the five are not listed in the table), |Spearman| with DMS on shared variants, 36 human DMS sets), percent10.31.22MutationAssessor
Livesey and Marsh 2025 DMS benchmark: pairwise |Spearman| win rate, direct assays onlyLivesey and Marsh 2025 human DMS benchmark set (36 proteins)Pairwise win rate (average pairwise win rate against all other VEPs, |Spearman| with DMS on shared variants, direct DMS assays only), percent59.317.5MutationAssessor
Livesey and Marsh 2025 DMS benchmark: pairwise |Spearman| win rate, indirect assays onlyLivesey and Marsh 2025 human DMS benchmark set (36 proteins)Pairwise win rate (average pairwise win rate against all other VEPs, |Spearman| with DMS on shared variants, indirect DMS assays only), percent58.321.7MutationAssessor
Livesey and Marsh 2025 DMS benchmark: pairwise |Spearman| win rate, SNV-accessible substitutions onlyLivesey and Marsh 2025 human DMS benchmark set (36 proteins)Pairwise win rate (average pairwise win rate against all other VEPs, |Spearman| with DMS on shared variants, SNV-accessible substitutions only), percent58.819.9MutationAssessor
Livesey and Marsh 2025 DMS benchmark: pairwise |Spearman| win rate, SNV-accessible substitutions with gaps filledLivesey and Marsh 2025 human DMS benchmark set (36 proteins)Pairwise win rate (average pairwise win rate against all other VEPs, |Spearman| with DMS, SNV-accessible substitutions, missing predictions filled with the VEP's most benign score), percent57.321MutationAssessor
Comparisons in Clinical likelihood ratios and balanced… (6)
ProtocolDatasetMetricMutationAssessorPrimateAIHigher
BRCA1 functional truth set, binary in silico calls (Cubuk et al. 2021)BRCA1 functional truth set: 1641 missense variants classed deleterious or tolerated by HAP1 cell survival after saturation genome editingBenignity likelihood ratio (Source 'negative likelihood ratio' (TNR/FNR): likelihood ratio for benignity of a tolerated call, BRCA1 functional truth set), unitless6.641.2MutationAssessor
BRCA1 functional truth set, binary in silico calls (Cubuk et al. 2021)BRCA1 functional truth set: 1641 missense variants classed deleterious or tolerated by HAP1 cell survival after saturation genome editingFalse negatives (Functionally deleterious BRCA1 variants called tolerated), count30283MutationAssessor
BRCA1 functional truth set, binary in silico calls (Cubuk et al. 2021)BRCA1 functional truth set: 1641 missense variants classed deleterious or tolerated by HAP1 cell survival after saturation genome editingLikelihood ratio (Positive likelihood ratio of a deleterious call (TPR/FPR), BRCA1 functional truth set), unitless2.12.92PrimateAI
BRCA2 functional truth set, binary in silico calls (Cubuk et al. 2021)BRCA2 functional truth set: 188 missense variants classed deleterious or tolerated by Homology-directed repairBenignity likelihood ratio (Source 'negative likelihood ratio' (TNR/FNR): likelihood ratio for benignity of a tolerated call, BRCA2 functional truth set), unitless9.881.23MutationAssessor
BRCA2 functional truth set, binary in silico calls (Cubuk et al. 2021)BRCA2 functional truth set: 188 missense variants classed deleterious or tolerated by Homology-directed repairFalse negatives (Functionally deleterious BRCA2 variants called tolerated), count150MutationAssessor
BRCA2 functional truth set, binary in silico calls (Cubuk et al. 2021)BRCA2 functional truth set: 188 missense variants classed deleterious or tolerated by Homology-directed repairLikelihood ratio (Positive likelihood ratio of a deleterious call (TPR/FPR), BRCA2 functional truth set), unitless1.166.78PrimateAI
Comparisons in An openly licensed benchmark… (6)
ProtocolDatasetMetricMutationAssessorPrimateAIHigher
CGC missense oncogenicity, all genes (Lee 2026, OncoCal tool_performance.tsv)Missense variants in 768 Cancer Gene Census genes with open oncogenicity labelsAUROC (CGC missense, all genes), unitless0.7940.797Tie
CGC missense oncogenicity, all genes (Lee 2026, OncoCal tool_performance.tsv)Missense variants in 768 Cancer Gene Census genes with open oncogenicity labelsAverage precision (CGC missense, all genes), unitless0.5540.508MutationAssessor
CGC missense oncogenicity, oncogenes (gain of function) (Lee 2026, OncoCal tool_performance.tsv)Missense variants in 768 Cancer Gene Census genes with open oncogenicity labelsAUROC (CGC missense, oncogenes (gain of function)), unitless0.7690.817PrimateAI
CGC missense oncogenicity, oncogenes (gain of function) (Lee 2026, OncoCal tool_performance.tsv)Missense variants in 768 Cancer Gene Census genes with open oncogenicity labelsAverage precision (CGC missense, oncogenes (gain of function)), unitless0.6040.618PrimateAI
CGC missense oncogenicity, tumour suppressors (loss of function) (Lee 2026, OncoCal tool_performance.tsv)Missense variants in 768 Cancer Gene Census genes with open oncogenicity labelsAUROC (CGC missense, tumour suppressors (loss of function)), unitless0.8240.872PrimateAI
CGC missense oncogenicity, tumour suppressors (loss of function) (Lee 2026, OncoCal tool_performance.tsv)Missense variants in 768 Cancer Gene Census genes with open oncogenicity labelsAverage precision (CGC missense, tumour suppressors (loss of function)), unitless0.5350.505MutationAssessor
Comparisons in ProteinGym (1)
ProtocolDatasetMetricMutationAssessorPrimateAIHigher
ProteinGym clinical substitutions, zero-shot, average AUC (commit 144fe22)ProteinGym clinical substitution benchmark at commit 144fe22AUROC (ProteinGym clinical substitutions, zero-shot, Average_AUC over proteins), unitless0.8770.855MutationAssessor

What disagrees

studies
2
Compared
MutationAssessor, PrimateAI
Study that disagrees
Other studies
Rule
Two-study conflict; unknown
Model identity
Matched across studies partly through links not yet reviewed

Follow-up

Current state: Not yet reviewed.

No one has reviewed this finding yet.

How to help resolve this

If you know why these studies disagree, for example a different population, metric or preprocessing step, or you have a corrected value, tell us through the contribution form. A reviewer checks every submission against its source before anything changes, and the step is added to the follow-up above.

Finding discrepancy-d353a619cccd3e87 · All contested results · How findings are made and followed up