Table 3 Quality parameters of IBS-M and IBD-M × F in different simulation scenarios

From: Estimating parent-specific QTL effects through cumulating linked identity-by-state SNP effects in multiparental populations

   

No noise

Medium noise

High noise

IBS-M

QTL detection powera

1 QTL

1.00

0.48

0.01

  

3 QTLs

0.99

0.78

0.37

  

8 QTLs

0.91

0.74

0.45

  

Mean

0.94

0.72

0.39

 

False positivesb

1 QTL

0.63

0.75

0.83

  

3 QTLs

0.54

0.57

0.73

  

8 QTLs

0.46

0.48

0.55

  

Mean

0.55

0.60

0.70

 

Prediction abilityc

1 QTL

0.65

0.13

0.03

  

3 QTLs

0.64

0.26

0.08

  

8 QTLs

0.55

0.28

0.08

  

Mean

0.61

0.22

0.07

IBD-M × F

QTL detection power

1 QTL

1.00

1.00

0.07

  

3 QTLs

1.00

1.00

0.34

  

8 QTLs

1.00

0.96

0.00

  

Mean

1.00

0.98

0.09

 

False positives

1 QTL

0.31

0.00

0.00

  

3 QTLs

0.22

0.00

0.00

  

8 QTLs

0.07

0.00

n/a

  

Mean

0.20

0.00

0.00

 

Prediction ability

1 QTL

0.97

0.35

0.00

  

3 QTLs

0.96

0.54

0.13

  

8 QTLs

0.95

0.68

0.00

  

Mean

0.96

0.52

0.04

  1. Abbreviations: IBD-M × F, identical-by-descent marker-by-family effects; IBS-M, identical-by-state marker main effects; QTL, quantitative trait locus.
  2. All values are averaged across 100 simulated cross-validation runs.
  3. aQTL detection power is defined as the model’s ability to precisely detect the simulated QTL within a 5 cM window surrounding the true position.
  4. bFalse positive associations are all detected significant associations that were outside the interval of cumulated single-nucleotide polymorphisms (SNPs; 26 cM) surrounding a QTL.
  5. cPrediction ability represents the correlation coefficient of predicted phenotypes (based on SNP effects obtained in the training set) and observed phenotypes in the test set.
Back to article page