Skip to main content

Table 7 RF-based prediction error rates of SNP sets from functional module-based filtration and SNP combinations with various thresholds from the WTCCC T2D dataset.

From: Finding type 2 diabetes causal single nucleotide polymorphism combinations and functional modules from genome-wide association data

Dataset Functional Module Description Number of SNPs Error Rate Number of Selected SNPs Error Rate with Variable Selection
MIR CACTGCC,MIR-34A, MIR-34C,MIR-449 1876 39.02% 6 32.12%
TF V$NKX25_02 1678 38.56% 59 33.06%
MIR ACTTTAT,MIR-142-5P 1572 38.32% 35 33.42%
Average Average of results from 66 functional modules 1614.86 39.18% 32.24 36.98%
Random
Average
Average of results from 66 SNP sets consisting of randomly selected 1615 SNPs 1615 38.77% 113.17 36.57%
Combination Top 5 functional modules 7590 38.96% 36 32.78%
Combination Top 66 functional modules 25663 38.66% 62 17.43%
p-value based Top 1 SNP 1 33.93% 1 33.93%
p-value based Top 10 SNPs 10 27.19% 10 27.19%
p-value based SNPs with Bonferroni threshold 82 11.76% 82 11.76%
p-value based SNPs with p-value < 0.01 854 14.79% 114 11.70%
p-value based All SNPs 42798 37.02% 83 11.47%