hall-lab / fastqtl Goto Github PK

Can you please explain in detail the meaning of the following section of the paper
2.2 Finding a candidate QTL per phenotype
For simplicity, we will focus on a single molecular phenotype P quantified in a set of N samples. Let G be the set of genotype dosages at L variant sites located within a cis-window of 6 W Mb of the genomic location of P. To discover the best candidate QTL for P, FastQTL measures Pearson product-moment correlation coefficients between P and all L variants in G, stores the most strongly correlated variant q [ G as candidate QTL, and assesses its significance by calculating a nominal P-value pn with standard significance tests for Pearson correlation. How can this be affected by a test where only one SNP and its LD proxies (r=0.8) is tested. Thanks!

Hi there,

Thanks for making the fastqtl tool. I have been trying to test it for some of my data.

I realised that it requires the sample count in the phenotype bed and the genotype vcf to be exactly the same, as shown here:

and throws an error otherwise.

I was wondering if this could be made slightly more friendly, for example, using only intersecting samples for further analysis.

It is not rare having slightly different numbers of samples in the genotype file and the phenotype file.

Thanks,
Zhihao

Hi, would you please upload an example VCF file that contains exactly what the program is looking for when it is reading an SV from LUMPY?

So far, I have run LUMPY on my samples, sorted (lsort) and merged (lmerge) using svtools. Once I sort and merge, there is no longer a separate column for each sample genotype. Instead, the sample ID is in column 8 in the ";" separated list.

Does your modified code handle this properly?

I tried running your modified program using one sample genotype per column (as fastQTL traditionally handles SNVs), but it returned NAs for the result. I assume this is because fastQTL failed to properly read the VCF with SVs in it.

An example VCF would really clear up a lot of confusion for me.

Thanks,
Alex

Which GSL version are you using? I have problems with permutation using GSL 2.4.

Edit:

Apparently, the issue is not with GSL but perhaps with my data. The error only occurs with permutation:

Processing gene [OR4F5]

• Number of variants in cis = 458
• Best correlation = -0.9987
• Number of permutations = 69 / 1000
  gsl: simplex2.c:372: ERROR: non-finite function value encountered
  Default GSL error handler invoked.
  Aborted

I've tried to run this on Redhat and Ubuntu, using FastQTL-2.184.linux.tgz binary and by compiling myself, as well as the GTEx eQTL Docker image. It always fails at the same point. Any ideas on what could be wrong and how to proceed?

Thank you!

Best,

Heini