← All datasets

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 ↗

Mean cross-validation scores for this dataset. Each score comes from the requested feature- and sample-cap cell, and failed fits use the constant predictor.

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

All metric values at this operating point
ModelMacro-F1MCCBalanced accuracyROC-AUC
RealTabPFN v20.97060.95920.97290.9993
RealTabPFN 2.50.96910.95860.96380.9952
TabPFN 30.9690.95850.96370.9963
TabPFN Wide 5k (ne3) ‡0.96220.95260.95270.9964
TabPFN Wide (8k) ‡0.96040.95260.94330.998
TabFM0.95780.95370.95620.9835
CatBoost0.9540.94650.94150.9942
Extra Trees0.94140.93160.95480.9969
Random Forest0.93390.91910.93820.9951
Logistic Regression0.92870.92330.92510.9932
TabDPT †0.92390.91790.91410.9737
MLP0.90810.89760.93450.994
KNN0.90280.89220.92930.989
AutoGluon0.90150.90510.89180.9646
TabM0.8950.88680.9220.9903
MITRA0.88930.89890.88060.986
XGBoost0.88490.89250.87350.9909
RealMLP0.87960.9040.86910.9968
LightGBM0.82130.86180.79170.9941
TabICL v20.248500.33330.5