Definition of genomic risk loci

[Siyu Pan]

Hi,

    Thank you for your tool, it has helped me with a lot of bioinformatics analysis.

    I am a postgraduate student. Recently, I am conducting bioinformatics analysis on a local computer and I will refer to the FUMA definition of locus. I have checked the FUMA platform manual but still do not understand.

  1）Is the definition of LD block in FUMA a reference to LDetect and is it the same as GWASATLAS (4.Definition of lead SNPs and risk loci)?

   In GWASATLAS, LD block was defined by the SNPs (P-value < 0.05) that are in LD with the independent significant SNP (r2 ≥ 0.6).

  2）What are the parameters of LD clump, --clump-p2, --clump-r2,--clump-kb?

ind.sig.snp       
plink --clump xx --clump-p1 5e-8?  --clump-p2, 0.01? --clump-r2 0.6 --clump-kb 1000?
lead snp
plink --clump xx --clump-p1 5e-8?  --clump-p2, 0.01? --clump-r2 0.1 --clump-kb 1000?
   3) The definition of the RISK locus is a complex task and I didn't find the source code on github (FUMA). Can you provide the code for this part if possible?

Looking forward to hearing from you.

[Marijn Schipper]

Hi, 

The definition of an LD block is indeed the same as the GWASATLAS.
You can find a better overview at https://fuma.ctglab.nl/tutorial#snp2gene under the heading " Risk loci and lead SNPs".

The code for generating risk loci is located in https://github.com/Kyoko-wtnb/FUMA-webapp/blob/master/storage/scripts/getLD.py .
This python script first performs the clumping of ind.sig.snps and afterwards lead snps. 

Afterwards, if two LD blocks which contain snps which are in LD with an ind.sig.snp from both blocks are within 250kb from eachoter, they are merged into a single risk locus.

I hope this answers your questions.

Cheers,
Marijn

Op donderdag 21 juli 2022 om 05:24:13 UTC+2 schreef siyupa...@gmail.com: