Skip to main content
Fig. 1 | BioData Mining

Fig. 1

From: r2VIM: A new variable selection method for random forests in genome-wide association studies

Fig. 1

Heatmaps of type I error of single SNPs in simulation study 1. Shown are results for logistic regression and r2VIM with several factors in the different scenarios (different sample sizes and mtry parameters in RF). Columns correspond to SNPs that are selected in at least one approach and are ordered by chromosomal position. Type I error is color-coded in gray with white and black denoting 0 and 1, respectively. In addition, LD information is shown at the top with SNPs in high (r 2 > 0.8) and moderate LD (0.3 < r 2 ≤ 0.8) colored in red and yellow

Back to article page
\