Figure 6.

Evaluation of the modeling for three OATH hits. The top four rows represent $–{\text{log}}_{10}\left(p\right),$ ${\lambda }_{\text{GC}},$ $\beta ,$ and ${\sigma }_{\beta },$ estimated using 32 possible OATH models given five covariates for these three SNPs. $–{\log }_{10}\left(p\right),$ ${\lambda }_{\text{GC}},$ $\beta ,$ and ${\sigma }_{\beta }$ are in ascending order according to their $–{\log }_{10}\left(p\right).$ In the top row, the corresponding OATH models are denoted by colored squares, indicating the inclusion or exclusion of covariates; there are two bars, gray and pink, in each cluster, representing $–{\log }_{10}\left(p\right)$ with or without adjustment for ${\lambda }_{\text{GC}}.$ An asterisk indicates that this (SNP) is significant under the corresponding model without adjustment for ${\lambda }_{\text{GC}};$ three asterisks indicate that this SNP is significant under the corresponding model, both with and without adjustment for ${\lambda }_{\text{GC}}.$ OATH, Open GWAS algoriTHm; SNP, single nucleotide polymorphism.