Skip to main content
Fig. 1 | BioData Mining

Fig. 1

From: Transition-transversion encoding and genetic relationship metric in ReliefF feature selection improves pathway enrichment in GWAS

Fig. 1

Overall evaluation strategy. (1) Preprocess the GWAS data by minor allele frequency and linkage disequilibrium filtering, (2) Apply feature selection algorithms (purple, right panel A-D), (3) Map top SNPs from methods A-D to gene symbols, and (4) Map genes to the target pathways to compare enrichment. The feature selection algorithms (right panel) are (A) ReliefF with combinations of nearest-neighbor and diff metrics (additional details in Fig. 2), (B) Lasso penalized regression with principal component correction, (C) Random Forest permutation importance, and (D) random sets of genes to assess pathway size bias

Back to article page
\