1
$\begingroup$

For QTL analysis in mice GEMMA was used to get P values ("p_lrt" column) for SNPs. GEMMA output (...assoc.txt) file excerpt:

chr  rs           ps       n_miss  allele1  allele0  af     logl_H1        l_mle         p_lrt  
1    UNC6         3010274  45      T        C        0.753  -1.871575e+03  9.092954e+00  4.497885e-01  
1    JAX00240613  3323400  12      T        C        0.766  -1.871777e+03  9.096464e+00  6.822045e-01  
...

Then, drawing a Manhattan plot with qqman R package applies default suggestive/genomewide cutoffs. But if I get it right, P < 5×10^(−8) is a commonly used genomewide threshold Ref1, Ref2.

Then I have to list the relevant coding/noncoding regions linked to these significant SNP (with P < 5×10^(−8)).
How can I do this?

I found a fantastic tool called FUMA
but it seems for GWAS in humans not for QTL analysis.

Update:
In Karl Browman's presentations: pdf (p. 13); pdf (p. 23(=28), 29(=34), 30(=35)).

Improve this question
$\endgroup$
2
  • 2
    $\begingroup$ GEMMA should give you confidence/credible intervals no? Also please expand on what you mean by ‘gene candidates’, that’s a bit of a vague term. $\endgroup$
    – user438383
    Commented Oct 11, 2022 at 15:16
  • $\begingroup$ thanks for the comment. I've edited the question. $\endgroup$
    – abc
    Commented Oct 13, 2022 at 11:29

1 Answer 1

Reset to default
2
$\begingroup$

I would suggest seeing what others have done when using GEMMA. For example, you can see this paper's methods:

We converted p-values to LOD scores and used a 1.5-LOD support interval to approximate a critical region around each associated region, which enabled us to systematically identify overlap with eQTLs.

They further discuss this approach in the supplementary note:

Instead we converted p-values to LOD scores and used a 1.5-LOD support interval to approximate a critical region around each association. The LOD drop approach provides a quick, straightforward way to gauge mapping precision and systematically identify overlap between eQTL genes and candidate QTGs; however, it does not correspond to a specific confidence interval (e.g. 95% confidence interval).

I would suggest trying something like this to estimate such a "critical region". For a discussion of the relationship between LOD and p-values, a quick search yielded this paper.

Improve this answer
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .