Supplemental Material for Tyler et al. 2019

Fig S1 QQ plots of p values To test for population structure in this patient cohort, we performed marker regression tests for each phenotype using sex as a covariate. We plotted the quantile normalized p values against the theoretical normal distribution. If population structure were affecting the SNP associations, we would expect to see systematic deviation from the line y = x in these plots. Rather we see that the distribution almost perfectly matches the theoretical distribution. The only deviations are a few significant p values in the HLA region for the ACA phenotype. We concluded that there was no population structure in this population and did not use a correction for population structure in our analysis.

Fig S2. Functional enrichment of co-expressed genes We searched across multiple subsets of data sets in the Gene Expression Omnibus (GEO) for genes significantly co-expressed with either RBMS3 and WNT5A or RBMS3 and MSI2. This series of figures reports all terms significantly associated with each pair in each data subset. Bars show the −log10(p) of the enrichment. Blue highlighting indicates a term matching one of the key words we used to identify SSc-related terms.