Skip to main content

Table 5 Accuracies 1 obtained for the PCR 2 and GREML 3 models

From: A comparison of principal component regression and genomic REML for genomic prediction across populations

Test

Trait

GREML

PCR_eigen

PCR_ss

SSPCR_eigen

SSPCR_ss

   

5-fold

stratified

5-fold

stratified

5-fold

stratified

5-fold

stratified

UK_1

Milk

0.26*

0.21 (249)

0.16 (204)

0.17 (142)

NA4 (0)

0.10 (71)

NA (0)

0.13 (821)

0.22 (537)

 

Fat

0.07

−0.08 (35)

−0.01 (103)

0.02 (95)

0.11 (1)

−0.03 (40)

NA (0)

0.07 (859)

0.12 (614)

 

Protein

−0.01

−0.05 (121)

−0.09 (69)

−0.02 (91)

NA (0)

−0.13 (91)

NA (0)

0.03 (910)

0.05 (570)

UK_2

Milk

0.10

0.12 (233)

0.12 (217)

0.05 (342)

0.14 (13)

0.12 (89)

0.13 (10)

0.04 (941)

0.03 (527)

 

Fat

−0.02

−0.13 (96)

−0.12 (37)

−0.11 (35)

−0.09 (28)

−0.08 (33)

−0.08 (27)

0.03 (909)

0.05 (523)

 

Protein

0.01

−0.03 (89)

−0.06 (46)

−0.11 (91)

0.06 (7)

0.04 (83)

0.02 (10)

−0.06 (780)

−0.01 (505)

SWE

Milk

0.16

0.15 (181)

0.14 (187)

0.12 (184)

0.04 (19)

0.03 (64)

0.04 (15)

0.09 (895)

0.07 (529)

 

Fat

0.09

0.10 (162)

0.10 (146)

0.04 (101)

0.07 (713)

0.07 (57)

−0.02 (1)

0.16 (859)

0.04 (680)

 

Protein

0.06

0.00 (92)

−0.04 (48)

0.01 (82)

−0.06 (29)

−0.10 (17)

−0.10 (16)

0.10 (1025)

0.10 (574)

IRL

Milk

0.06

0.05 (277)

−0.10 (48)

−0.03 (207)

−0.08 (37)

−0.04 (35)

−0.13 (8)

0.14 (717)

0.13 (389)

 

Fat

0.08

0.06 (121)

0.04 (51)

0.02 (100)

0.08 (52)

0.07 (37)

0.02 (14)

0.07 (712)

0.06 (315)

 

Protein

0.04

0.05 (127)

0.04 (145)

0.00 (35)

−0.04 (173)

−0.05 (37)

−0.12 (12)

0.12 (759)

0.11 (421)

NLD

Milk

0.16

0.18 (50)

0.19 (28)

0.09 (55)

0.10 (7)

0.18 (25)

0.13 (2)

0.11 (95)

0.11 (95)

 

Fat

0.15

0.11 (47)

0.10 (99)

0.05 (75)

0.06 (7)

0.07 (39)

0.07 (7)

0.10 (197)

0.11 (196)

 

Protein

0.13

0.14 (31)

0.13 (34)

0.07 (42)

0.13 (7)

0.12 (28)

0.16 (2)

0.11 (48)

0.11 (48)

Average

Milk

0.15

0.14 (198)

0.10 (137)

0.08 (186)

0.05 (15)

0.08 (57)

0.04 (7)

0.09 (694)

0.02 (415)

 

Fat

0.07

0.01 (92)

0.02 (87)

0.00 (81)

0.05 (160)

0.02 (41)

0.00 (10)

0.07 (707)

0.04 (466)

 

Protein

0.05

0.02 (92)

0.00 (68)

−0.01 (68)

0.02 (43)

−0.02 (51)

−0.01 (8)

0.02 (704)

0.02 (424)

SD

Milk

0.08

0.06 (90)

0.12 (91)

0.08 (105)

0.10 (14)

0.08 (26)

0.12 (6)

0.04 (345)

0.07 (189)

 

Fat

0.06

0.11 (53)

0.09 (44)

0.07 (28)

0.08 (310)

0.07 (9)

0.06 (11)

0.05 (295)

0.04 (204)

 

Protein

0.05

0.08 (38)

0.09 (45)

0.07 (27)

0.09 (73)

0.10 (34)

0.13 (7)

0.08 (382)

0.05 (219)

  1. For the PCR models, the PC included (numbers are presented in brackets) were selected based on cross-validation in the reference population data, either in a 5-fold random or stratified split to select the optimum PCR model in respect to minimum mean squared error. Analyses were performed for three traits and five test populations.
  2. 1Accuracies were calculated as Pearson correlations between the predicted genomic breeding values and the adjusted phenotypes; 2selection of PCs was based either on the eigenvalues (eigen) or the regression sum of squares (ss); two different methods of applying principal component analysis, either separately for reference and test parts (PCR) or on the whole dataset (SSPCR), were compared; 3a REML based model with a genomic relationship matrix; 4all animals received the same prediction; *in italics the highest accuracies for each population and trait.
Back to article page

Genetics Selection Evolution

ISSN: 1297-9686