Graphs present the dependency of false positive counts given the number of selected best candidate interactions. Curves closer to lower-right corner of the graph indicate better performance. The axes are in logarithmic scale to emphasize the results for smaller numbers of best candidates.
We provide results for 12 simulated data sets and 5 experimental data sets.
light gray - theoretically best and worst possible performance curves
black solid - direct scoring
black dashed - scoring with two replication groups
black dotted - scoring with three replication groups
red solid - scoring with 10 bootstrap samples instead of data partitioning
red dashed - scoring with 20 bootstrap samples instead of data partitioning
red dotted - scoring with 50 bootstrap samples instead of data partitioning
yellow solid - scoring with 100 bootstrap samples instead of data partitioning
yellow dashed - scoring with 500 bootstrap samples instead of data partitioning
We followed the data synthesis by Ritchie et al. (2003). The simulated data sets were generated according to six two-SNP epistasis model. Unlike Ritchie et al. (2003), our data sets included multiple interactions, but such that each SNP was involved in interaction with at most one other SNP. Two different types of data sets with respect to the number of SNPs were crafted, each comprised 200 control and 200 disease samples:
syn1 (100 SNP data set) without noise
syn1 (100 SNP data set) with missing data noise
syn1 (100 SNP data set) with genotyping noise
syn1 (100 SNP data set) with phenocopies noise
syn1 (100 SNP data set) with genetic heterogeneity noise
syn1 (100 SNP data set) with all types of noise applied simultaneously
syn2 (500 SNP data set) without noise
syn2 (500 SNP data set) with missing data noise
syn2 (500 SNP data set) with genotyping noise
syn2 (500 SNP data set) with phenocopies noise
syn2 (500 SNP data set) with genetic heterogeneity noise
syn2 (500 SNP data set) with all types of noise applied simultaneously
Families with two individuals affected by autism spectrum disorders. Individuals were classified to affected or unaffected. Due to HFCC software constraints, only the first 2,000 SNPs were considered and a stratified sample of 500 individuals (292 affected, 208 unaffected) was used.
901 SNPs for each of the 121 cancerous samples and 87 controls.
1,189 SNPs for each of the 141 cancerous samples and 89 controls.
GSE7226-GPL2004
Platform designation GPL2004, comprising 102 samples from mentally retarded children and 213 controls from their unaffected siblings or parents. The first 2,000 SNPs were considered.
GSE7226-GPL2005
Platform designation GPL2005, comprising 103 samples from mentally retarded children and 210 controls from their unaffected siblings or parents. The first 2,000 SNPs were considered.
