Read on Twitter
Welcome all! We've added several hundred followers over the past few weeks, so as a quick intro, I use this account mainly to explore interesting issues related to the biological interpretation of GWAS
@SbotGwa from @andganna provides me with a steady diet of interesting material
@SbotGwa from @andganna provides me with a steady diet of interesting material
@SbotGwa alternates between GWAS from @uk_biobank and @FinnGen_FI.
Yesterday's Manhattan plot from FinnGen yielded a single hit for the trait "other and unspecified corneal deformities and disorders"
Let's dive in,
https://twitter.com/SbotGwa/status/1355137750803091460
Yesterday's Manhattan plot from FinnGen yielded a single hit for the trait "other and unspecified corneal deformities and disorders"
Let's dive in,
https://twitter.com/SbotGwa/status/1355137750803091460
I often say it's good to take the most significant association at a locus as we start to interpret it.
@FinnGen_FI uses a modified PheWeb server to show results.
Here is the PheWAS for this SNP.
Top association is Keratitis, inflammation of the cornea.
https://r4.finngen.fi/variant/2-227207329-T-C
@FinnGen_FI uses a modified PheWeb server to show results.
Here is the PheWAS for this SNP.
Top association is Keratitis, inflammation of the cornea.
https://r4.finngen.fi/variant/2-227207329-T-C
But actually strongest p is a poor surrogate for "biologically most important"
I made this complex plot to show the relationship between effect size, p-value and case numbers for the disease associations at this locus.
Dot size and numbers indicate case numbers which vary a lot
I made this complex plot to show the relationship between effect size, p-value and case numbers for the disease associations at this locus.
Dot size and numbers indicate case numbers which vary a lot
This relates to "clumping and splitting" of GWAS traits. The more narrow phenotype has a much larger effect size, but a weaker p-value because the smaller case counts reduces statistical power.
So what to make of "other and unspecified"?
Doesn't sound very specific...
So what to make of "other and unspecified"?
Doesn't sound very specific...
According to the @GWASCatalog this locus has been associated with numerous eye traits including corneal thickness, intraocular pressure and macular thickness.
https://www.ebi.ac.uk/gwas/genes/COL4A3
https://www.ebi.ac.uk/gwas/genes/COL4A3
The cornea of course is the clear covering at the front of the eye.
https://en.wikipedia.org/wiki/Cornea
https://www.mayoclinic.org/diseases-conditions/keratoconus/symptoms-causes/syc-20351352#dialogId8508157
https://en.wikipedia.org/wiki/Cornea
https://www.mayoclinic.org/diseases-conditions/keratoconus/symptoms-causes/syc-20351352#dialogId8508157
The lead SNP in this GWAS sits within an intron of COL4A3.
As a general rule, the gene closest to the lead SNP is usually the true causal gene, especially if the lead SNP is within the footprint of the gene. https://genetics.opentargets.org/variant/2_227207329_T_C
As a general rule, the gene closest to the lead SNP is usually the true causal gene, especially if the lead SNP is within the footprint of the gene. https://genetics.opentargets.org/variant/2_227207329_T_C
COL4A3 encodes a collagen protein which contributes to the structural integrity of the cornea.
Rare mutations in COL4A3 lead to keratoconus where the cornea thins and bulges outward forming a cone shape.
https://medlineplus.gov/genetics/gene/col4a3/#conditios
https://www.mayoclinic.org/diseases-conditions/keratoconus/symptoms-causes/syc-20351352#dialogId8508157
Rare mutations in COL4A3 lead to keratoconus where the cornea thins and bulges outward forming a cone shape.
https://medlineplus.gov/genetics/gene/col4a3/#conditios
https://www.mayoclinic.org/diseases-conditions/keratoconus/symptoms-causes/syc-20351352#dialogId8508157
So I'll wrap up here.
Not only can we conclude that COL4A3 is likely the causal gene at the locus, but also that the 330 "unspecified corneal deformities" likely represents a specific form of keratoconus.
(the image is collagen)
Not only can we conclude that COL4A3 is likely the causal gene at the locus, but also that the 330 "unspecified corneal deformities" likely represents a specific form of keratoconus.
(the image is collagen)
Surprisingly the @finngen GWAS of keratoconus is flat and shows no association at COL4A3 or anywhere else. Not really sure what that means.
http://r4.finngen.fi/pheno/H7_CORNEALDEFORM
http://r4.finngen.fi/pheno/H7_CORNEALDEFORM
