Table 3 Predictive abilities obtained with LASSO, the elastic net, BayesB, and RR-BLUP for three experimental data sets
Predictive ability
DataTraitnLASSOElastic netBayesBRR-BLUP
RiceFlowering time3590.46 ± 0.0130.52 ± 0.0110.59 ± 0.0060.59 ± 0.007
Plant height3830.71 ± 0.0050.73 ± 0.0030.76 ± 0.0020.76 ± 0.002
Panicle length3750.60 ± 0.0040.61 ± 0.0070.66 ± 0.0060.66 ± 0.006
Seed length3770.75 ± 0.0030.78 ± 0.0030.75 ± 0.0030.75 ± 0.003
WheatYield2540.44 ± 0.0100.43 ± 0.0060.51 ± 0.0060.51 ± 0.005
TKWa2540.48 ± 0.0100.48 ± 0.0110.55 ± 0.0040.55 ± 0.004
Days to heading2540.60 ± 0.0050.61 ± 0.0080.66 ± 0.0050.66 ± 0.004
ArabidopsisFlowering time1800.71 ± 0.0080.65 ± 0.0040.75 ± 0.0040.75 ± 0.004
Plant diameter1800.47 ± 0.0120.49 ± 0.0160.51 ± 0.0060.51 ± 0.007
FRIb1640.48 ± 0.0200.36 ± 0.0260.41 ± 0.0140.41 ± 0.013
Plant width1760.46 ± 0.0240.49 ± 0.0100.44 ± 0.0090.43 ± 0.009
  • Predictive abilities for all 11 traits were estimated using fivefold cross-validation and results were averaged over 10 replications ± SE.

  • a TKW, thousand-kernel weight.

  • b FRI, FRI gene expression.