Fig. 1
Interaction Analysis identifies 14 unique risk loci pairs. a Circos plot of genome-wide association and significant interaction results for the identified paired risk loci. The second outer most panel displays results from genome-wide association study on a Manhattan plot for autosomal variants on a log transformed scale (−log10) of 1 to 8.5. Negative log transformed interaction P values corresponding to each of the interaction pair is calculated from log linear transformed regression on the discovery set and is represented on an adjusted inflated scale of 9.3 to 10.2 in the second inner most panel. More than one unique variant pair combinations are present in the same interacting regions which are marked with their corresponding odds in this panel. Genome-wide significant paired loci are line-joined in the inner most panel based on their chromosomal positions (NCBI build 19 human genome). Annotations of single nucleotide polymorphisms to gene ids are displayed at the outer most panel. b Forest plot with embedded confidence intervals for each of the identified interaction pairs. Each pair indicates two interacting chromosomal locations with base pair information for the indexing loci. Paired variants annotated to the same indexing chromosomes are line joined. chr chromosome, BP base pair; OR, odds ratio; CI, confidence interval