Skip to main content

Table 4 Accuracies of prediction for all prediction scenarios in the study

From: Utility of whole-genome sequence data for across-breed genomic prediction

 

Across-breed, across-country

Across-breed, within-country

Within-breed, across-country

Reference

DH

NZJ

NZH

NZJ

NZH

DH

Validation

NZJ

DH

NZJ

NZH

DH

NZH

Scenarios with unselected markers (number of markers)

Full_seq (14,341,737)

0.08

0.02

0.09

0.06

0.41

0.51

HD (583,078)

0.10

0.05

0.07

− 0.04

0.43

0.55

50k (48,912)

0.06

0.05

0.13

− 0.04

0.42

0.52

Scenarios with unselected markers + selected markers (number of markers)

HD_Top (583,194

0.10

0.05

0.07

− 0.04

0.43

0.55

50k_Top (49,045)

0.07

0.06

0.13

− 0.03

0.43

0.53

Scenarios with only selected markers (number of markers)

Pval5 (59,828)

0.19

0.19

0.14

0.32

0.52

0.58

Pval7 (23,125)

0.01

0.15

0.07

0.35

0.46

0.47

COJO3 (1570)

0.20

0.22

0.14

0.31

0.45

0.52

COJO8 (360)

0.21

0.25

0.16

0.18

0.47

0.44

Top_markers (133)

0.23

0.18

0.08

0.27

0.41

0.47

SE*

0.04

0.01

0.04

0.03

0.01

0.01

  1. Predictions were carried out either across breed and across country, across breed and within country or within breed and across country. The populations are Dutch Holsteins (DH), New Zealand Holsteins (NZH) and New Zealand Jerseys (NZJ)
  2. The unselected marker sets are the whole-genome sequence (Full_seq), high-density markers (HD) and markers on the traditional 50k chip (50k). HD_Top and 50k_Top are scenarios in which some pre-selected markers from a meta-GWAS (Top_markers) are added to the HD and 50k markers respectively. Pval5 and Pval7 are marker sets pre-selected from a meta-GWAS based on their p values. The COJO scenarios are those containing markers that are assumed to be independently significant markers from a meta-GWAS at different significant levels. Top_markers contain the most significant markers in each QTL region from a meta-GWAS
  3. *Standard error (SE) of estimates, did not differ across the different sets of markers, provided the reference and validation populations remained the same