Skip to main content

Table 3 Performance of the different combinations of methods for optimizing training set (TRS) size and composition

From: Maximizing efficiency in sunflower breeding through historical data optimization

Trait

TRS Optimization Method

Average TRS size and gain in predictive ability both expressed as a percentage relative to the entire candidate set

type

Size

Composition

Test Set year 6

Test Set year 7

Globally

Mean

sd

TRS size

Mean

sd

TRS size

Mean

sd

TRS size

YLD

Genetic

Min_GRM

Avg_GRM_self

98.22

2.37

79.29

98.39

2.24

79.81

98.31

2.30

79.55

Min_GRM

Avg_GRM_MinMax

98.03

2.45

79.29

98.26

2.84

79.81

98.14

2.67

79.55

Min_GRM

Avg_GRM

96.82

2.44

79.29

97.01

3.28

79.81

96.92

2.92

79.55

Min_GRM

Min_GRM

97.64

3.29

79.29

95.44

3.22

79.81

96.54

3.46

79.55

Min_GRM

PCA_CDmean

98.67

2.80

79.29

97.94

2.26

79.81

98.31

2.56

79.55

Mixed

Min_GRM

PLS_CDmean

99.13

2.24

79.29

98.56

2.54

79.81

98.85

2.43

79.55

Tails_GEGVs_sd1

Tails_GEGVs

101.41

2.24

60.62

98.95

1.22

60.40

100.18

2.17

60.51

Manually set 60%

Tails_GEGVs

102.58

2.48

60.00

98.21

1.51

60.00

100.39

3.00

60.00

Phenotypic

Min_GRM

Tails

98.94

2.44

79.29

98.85

1.88

79.81

98.90

2.14

79.55

Min_GRM

Random

98.62

2.44

79.29

97.90

2.46

79.81

98.26

2.49

79.55

GM

Genetic

Min_GRM

Avg_GRM_self

98.80

1.56

79.29

98.86

2.12

79.81

98.83

1.89

79.55

Min_GRM

Avg_GRM_MinMax

98.76

1.48

79.29

98.25

2.48

79.81

98.51

2.10

79.55

Min_GRM

Avg_GRM

99.27

4.28

79.29

96.98

2.70

79.81

98.13

3.66

79.55

Min_GRM

Min_GRM

98.43

1.65

79.29

97.70

1.57

79.81

98.06

1.65

79.55

Min_GRM

PCA_CDmean

98.44

1.82

79.29

98.37

2.65

79.81

98.40

2.31

79.55

Mixed

Min_GRM

PLS_CDmean

98.25

1.50

79.29

98.21

2.37

79.81

98.23

2.02

79.55

Tails_GEGVs_sd1

Tails_GEGVs

95.11

2.80

56.19

93.85

3.50

59.36

94.48

3.23

57.77

Manually set 60%

Tails_GEGVs

96.56

1.23

60.00

95.21

3.14

60.00

95.89

2.53

60.00

Phenotypic

Min_GRM

Tails

98.68

1.07

79.29

100.37

1.57

79.81

99.52

1.61

79.55

Min_GRM

Random

98.56

1.59

79.29

98.70

2.84

79.81

98.63

2.36

79.55

OIL

Genetic

Min_GRM

Avg_GRM_self

98.62

1.53

79.29

98.84

2.28

79.81

98.73

1.98

79.55

Min_GRM

Avg_GRM_MinMax

98.71

1.57

79.29

98.91

2.36

79.81

98.81

2.04

79.55

Min_GRM

Avg_GRM

95.69

5.11

79.29

100.54

2.94

79.81

98.12

4.83

79.55

Min_GRM

Min_GRM

96.40

1.80

79.29

100.21

1.65

79.81

98.30

2.58

79.55

Min_GRM

PCA_CDmean

98.99

1.44

79.29

99.12

2.08

79.81

99.05

1.82

79.55

Mixed

Min_GRM

PLS_CDmean

98.51

1.58

79.29

99.00

2.15

79.81

98.76

1.93

79.55

Tails_GEGVs_sd1

Tails_GEGVs

94.26

2.86

62.03

96.38

2.49

62.07

95.32

2.87

62.05

Manually set 60%

Tails_GEGVs

94.34

2.37

60.00

96.42

2.21

60.00

95.38

2.51

60.00

Phenotypic

Min_GRM

Tails

98.90

1.11

79.29

99.62

1.62

79.81

99.26

1.44

79.55

Min_GRM

Random

98.18

1.66

79.29

98.74

2.34

79.81

98.46

2.07

79.55

  1. The optimized training sets for all traits were evaluated using 30 repetitions of gradient boosting machine model. For each test set, the average performance across the different candidate sets tested is displayed. Furthermore, the average for both test sets is in the “Globally” column. The performance values are expressed as a percentage of the predictive ability obtained using the entire candidate set to calibrate the models and the training set size is expressed as a percentage of the candidate set size
Back to article page

Plant Methods

ISSN: 1746-4811