Skip to main content

Table 1 Predictive accuracy measures in the literature for models for prediction of variants associated with complex traits

From: Assessing models for genetic prediction of complex traits: a comparison of visualization and quantitative methods

   

Predictive accuracy measures employed

 

Algorithm

Classifier

Area under ROC curve

Positive predictive value

Box plot

Histo-Gram

Violin plot

Mann–Whitney U/Wilcoxon Rank Sum test

Gagliano et al. 2014

Modified Elastic net

GWAS hits vs. non-hits

x

x

 

x

  

Iversen et al. 2014

Penalized logistic regression

GWAS hits vs. non-hits

x*

     

Kircher et al. 2014

Support Vector Machines

High-frequency human-derived alleles vs. simulated variants

x

   

x

x

Ritchie et al. 2014

Modified Random Forest

HGMD hits vs. non-hits

x

 

x

  

x

  1. *reports “Concordance index”, which is equivalent to the area under the ROC curve