Table 2. Cross-validation scenarios.

Set Name	Schema	Training Size	Validation Size	Marker Number
CVC0	C0$\mathrm{\circlearrowright }$	122	31	130
CVC2	C2$\mathrm{\circlearrowright }$	132	34	1951
CVT1	T1$\mathrm{\circlearrowright }$	136	35	3307
CVStrat	(C0,C2,T1) $\to$ (C0,C2,T1)	48-402	75	535
Prog	C0 $\to$ C2	179	168	535
Test	C0 $\to$T1	179	172	535

Summary of different cross-validation (CV) methods used in this study. Within-population schemes were tested in C0 (CV_C0), C2 (CV_C2), and in the test population (CV_T1). Across population testing included predicting progeny from C0 (Prog) and T1 from C0 (Test). A stratified approach (CV_Strat) was used to determine the role of population size on prediction accuracy.