Pooled sample for GWAS power

[laurence_13]

Hi,

I'm very new to the forum, if you have any suggestions, please share them.

So i'm working with honey bee to identify QTL associated to a resistance trait (quantitative trait) using GWAS (by calling SNPs).
Honey bee queens are diploid and males progeniture inherit 1/2 of their genes (so they are haploid).
We want to sequence the queens, but can't use them, so we sampled the males progeny. I made 4 pools of 6 males for each queen.

My question is this: Is it better to do all my pipeline to SNP calling on each of the 4 pools individually before joining the results together, or should i joint the pools raw reads before starting my pipeline to enhance statistical power?

My logic is that if I don't initially joint them, I can later compare allele frequency between each replicate to insure that rare variants are not eliminated.

thanks

[gringer]

My usual approach is to do the mapping separately, and do variant calling in parallel with samtools mpileup in multi-BAM mode (i.e. using more than one BAM file on the command line). This increases the chance that a rare/odd variant will be picked up, and makes sure that all variants are called in all samples.