Fig. 3

A Flowchart showing our 3′ UTR microRNA-binding site exploration pipeline using 3 publically available datasets and summary statistics provided by IMSGC (2019). 3′UTR variant collation was performed separately from microRNA variant collation. We obtained variants from dbSNP v151 and integrated these into the 3′UTR binding sites predicted by RNA22 v2.0 and TargetScan v7.0 (see Methods) (B). Schematic showing the predicted effects of rs6742 on microRNA-binding ability to 3′UTR in SLC2A4RG. The C allele is expected to bind to 6 microRNAs differently to the T allele. Overall, we expect that the C allele is under stronger regulation than the T allele. C Venn Diagram showing the overlap between GWAS independent SNPs and our collection of 3′UTR SNPs which are in 3′UTR binding sites. Independent SNPs were identified through FUMA and a list of suggestive effects provided by the IMSGC. D Among the 19 independent SNPs from C, we tested the microRNA-binding ability of the 3′UTR binding sites containing 8 SNPs. Among these, 6 SNPs were found to cause microRNA-binding site changes in their 3′UTR sites. Here, we show the number of microRNAs that bind to the alternative and reference versions of the 3′UTR sequences, as well as the microRNAs that bind differently. The source column highlights which GWAS independent list the SNP has been output from. E LocusZoom regional plot showing the 3′UTR SNP rs2587100, which is independent, weakly suggestive, causes changes in microRNA-binding ability of BCL2L13 and is an eQTL for BCL2L13 in general monocytes and MS patient monocytes. This is our only candidate SNP which has MS patient specific eQTL evidence. No other SNPs in this region were prioritised among the genome-wide IMSGC SNPs. The highlighted intronic SNP rs9618043 (CECR2) is among the non-replicated SNPs (NR) from the IMSGC’s prioritised effects within this region (IMSGC Additional file 1: Table S6), while rs9618040 is not among the prioritised effects (intron CECR2)