Regional association plots
Three exemplar loci, one per major result. Hover any point for position, p-value, and LD to the lead variant; toggle tracks with the checkboxes.
How to read a regional plot
What this data is
Every tested variant in a window around one locus. Position runs along the x axis; association strength (larger means a smaller p-value) runs up the y axis. The top track is the schizophrenia GWAS; the tracks below it are molecular QTL datasets over the same window, with gene positions along the bottom.
How to read it
Points are colored by LD with the lead variant, from red (near-perfect correlation) down through the legend’s bins; the lavender diamond is the lead variant itself, and gray points are uncorrelated or absent from the LD panel. Colocalization appears as matching skylines: the same colored points rise to a peak in two tracks. Two signals side by side in different colors are the visual signature of distinct causal variants.
What to take away, and what not to
- FURIN shows what true sharing looks like; the rs9607782 case shows how a single-variant test can be fooled when a locus carries two signals; the rs6010045 case shows a methylation chain nominating a gene without direct expression evidence.
- Visual similarity is suggestive only. The numbers elsewhere on this site come from the formal colocalization test, not from eyeballing skylines.
The benchmark schizophrenia variant-to-gene case. The GWAS signal at rs4702 and the FURIN eQTL in MetaBrain cortex share the same lead variant and LD structure; coloc assigns PP4 = 1.00.
Loading…