Table 1.

Similarity of random and balanced subsets with populations and comparison of population-level predictive performance with that estimated with sampled subsets

Prediction algorithm	Yeast
	Population	Random subsets	Balanced subsets
M1	0.71 ± 0.02	0.70 ± 0.02	0.42 ± 0.02
M2	0.67 ± 0.02	0.66 ± 0.02	0.52 ± 0.02
M3	0.57 ± 0.01	0.57 ± 0.02	0.53 ± 0.02
M4	0.71 ± 0.02	0.71 ± 0.02	0.62 ± 0.02
Similarity to population		0.00 ± 0.03	−0.71 ± 0.01
Prediction algorithm	Human
	Population	Random subsets	Balanced subsets
M1	0.72 ± 0.01	0.72 ± 0.01	0.45 ± 0.01
M2	0.67 ± 0.01	0.67 ± 0.01	0.49 ± 0.02
M3	0.58 ± 0.01	0.58 ± 0.02	0.51 ± 0.02
M4	0.72 ± 0.01	0.71 ± 0.01	0.63 ± 0.02
Similarity to population		0.02 ± 0.01	−1.00 ± 0.00

Similarity is reported in the form of the average correlation coefficient ± the standard deviation. Predictive performance is reported in the form of the average AUC ± the standard deviation.