FINEMAP output as part of FLAMES pipeline

[Jamee Berg]

Hi, Tanya,

Thank you so much for the FLAMES pipeline.

1. Is there a way to view the FINEMAP output, which is an intermediate step in FLAMES?
2. Does the software check for Allele Flips and ambiguous SNPS (A/T, G/C) before running FINEMAP?

Thank you so much!

Jamee

[Tanya Phung]

Hi Jamee, 

To answer your questions: 

2. If you run FLAMES with the processed gwas sumstat processed by FUMA SNP2GENE, then in this file, the chr:pos:a1:a2 is checked against dbSNPv146. 

- If you run FLAMES by supplying your own gwas sumstat, then FUMA does not check for alleles.

1. In the FUMA implementation of FLAMES, the finemapping step is run with polyfun which estimates prior causal probabilities for SNPs, which can then be used by fine-mapping methods like SuSiE or FINEMAP (source: https://github.com/omerwe/polyfun). You can also check this tutorial on how polyfun is implemented as part of the FLAMES workflow: https://github.com/tanyaphung/FLAMES/blob/add_tutorial/tutorial/tutorial.md. 

So polyfun returns the causal probability for SNPs, and the file looks like this per genomic risk locus:
SNP     PIP
2:49286400:A:G  0.0801448
2:49299046:T:C  0.0618813
2:49287066:A:G  0.0494603
2:49296396:T:C  0.0472717
2:49291893:A:T  0.0319731
2:49292341:A:G  0.0283623
2:49293148:A:G  0.0260545

Is this something you are interested in being able to download in the FLAMES module? 

I can add it as an option to download. In the meantime, if you need it sooner, you can email me with the jobID and I can send it to you. 

Best,
Tanya