Strategies to improve genomic predictions for 35 duck carcass traits in an F2 population

Table 2 The predictive reliability of (genomic) breeding values for duck carcass traits using different strategies

Methods	Weight traits	Length traits	Percentage traits	All 35 traits
BLUP	0.318	0.205	0.250	0.269
Sum 2p(1 − p) GBLUP	0.395	0.240	0.298	0.327
Independent 2p(1 − p) GBLUP	0.400	0.240	0.300	0.330
*Sum true variance GBLUP	0.399	0.243	0.302	0.331
*Independent true variance GBLUP	0.404	0.243	0.304	0.334
BayesB	0.468	0.274	0.304	0.364
BayesCπ	0.405	0.244	0.305	0.334
BayesR	0.360	0.205	0.285	0.301
BayesS	0.407	0.244	0.308	0.337
BayesN	0.507	0.301	0.307	0.386

Data showing the average of predictive reliability in each trait group. The predictive reliability is computed using 5-fold cross-validation. The true variance method, denoted with an asterisk (*), is a newly proposed method for improving GRM in the GBLUP model

ISSN: 2049-1891