Genomic prediction

Switchgrass Height Genomic Prediction

This switchgrass panel pairs SNP genotype profiles with plant-height measurements for individual plant lines. The regression task predicts height from genetic markers using the data assembled for Azodi and colleagues' comparison of genomic-prediction methods.

RegressionLicence: CC0-1.0gp-switchgrass-HTTarget: HTgp ↗

At this operating point

Model comparison

p=10,000, n=100 · Lower RMSE 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	RMSE	MAE	R²
Extra Trees	0.8166	0.6414	0.3256
Random Forest	0.8214	0.6447	0.3172
Extra Trees (tuned)	0.8233	0.6496	0.3131
Logistic Regression (tuned)	0.8234	0.6531	0.3116
Random Forest (tuned)	0.8259	0.6499	0.3093
Logistic Regression	0.8262	0.6558	0.3063
TabPFN 3.5	0.8265	0.65	0.3059
TabFM	0.8269	0.661	0.3051
RealTabPFN 2.5	0.8276	0.6447	0.3043
GeneICL	0.8294	0.6594	0.3009
Kumo Tabular (M)	0.8325	0.6637	0.2939
RealTabPFN v2	0.8331	0.6525	0.2961
TabM	0.8606	0.6782	0.2488
TabDPT	0.8621	0.6875	0.2477
RealMLP	0.8624	0.6867	0.2467
AutoGluon	0.8684	0.6952	0.2366
MITRA	0.8831	0.7003	0.2099
MLP	0.8887	0.7081	0.1978
Causilo	0.9026	0.7041	0.1682
LimiX2	0.9173	0.7278	0.15
XGBoost (tuned)	0.9195	0.7319	0.1467
XGBoost	0.9483	0.75	0.09362
MLP (tuned)	1.158	0.977	-0.5093