Metagenomics

Human Microbiome Environment Metagenome

Microbial functional profiles from Human Microbiome Project samples collected across several body sites in healthy adults. The benchmark uses MGnify InterPro abundances to predict the recorded isolation source of each sample.

Binary classificationLicence: PublicMGYS00001056Target: isolation sourcemgnify ↗

At this operating point

Model comparison

p=10,000, n=100 · Higher Macro-F1 is better

Random Forest is marked with a diamond. Models are ordered by the selected metric.

Across sample budgets

Learning curves

Models

All metric values at this operating point

Model	Macro-F1	MCC	Balanced accuracy	ROC-AUC
Kumo Tabular (M)	0.9747	0.9704	0.9708	0.9999
Causilo	0.9731	0.9705	0.9614	0.9986
RealTabPFN v2	0.9706	0.9592	0.9729	0.9993
LimiX2	0.9696	0.9651	0.9784	0.9985
RealTabPFN 2.5	0.9691	0.9586	0.9638	0.9952
TabPFN 3.5	0.965	0.9584	0.9562	0.9953
TabPFN Wide 5k (ne3)	0.9622	0.9526	0.9527	0.9964
TabFM	0.9578	0.9537	0.9562	0.9835
Extra Trees	0.9414	0.9316	0.9548	0.9969
Random Forest	0.9339	0.9191	0.9382	0.9951
Random Forest (tuned)	0.9299	0.924	0.9249	0.9958
Logistic Regression	0.9287	0.9233	0.9251	0.9932
Logistic Regression (tuned)	0.9255	0.9174	0.9323	0.9931
TabDPT	0.9239	0.9179	0.9141	0.9737
Extra Trees (tuned)	0.919	0.9127	0.9215	0.9941
MLP (tuned)	0.9115	0.9139	0.9217	0.9951
XGBoost (tuned)	0.9105	0.912	0.8975	0.9947
MLP	0.9081	0.8976	0.9345	0.994
AutoGluon	0.9015	0.9051	0.8918	0.9646
TabM	0.895	0.8868	0.922	0.9903
MITRA	0.8893	0.8989	0.8806	0.986
XGBoost	0.8849	0.8925	0.8735	0.9909
RealMLP	0.8796	0.904	0.8691	0.9968
GeneICL	0.8219	0.8061	0.9182	0.9972