Statistical analysis, such as SNPs association, of pooled-seq data

Question

[TH Yeh]

Hi guys, I'm curious that is it possible to do GWAS/ SNPs association on pooled sequencing data without a control? To my understanding, GWAS/ SNPs association is an analysis comparing SNPs between experiment and control group on every position, so without proper control, every SNP in the experiment group will be significant?

For example, The data I have:

• Experiment-- Three sequencing files, each contains a different number of pooled random EMS-mutagenized drug-resistant flies.
• A stander Drosophila reference .fasta file. The data I don't have:
• Control-- Pooled random EMS-mutagenized drug-non-resistant flies. or simply a random EMS-mutagenized group without drug selection.

I've been calling variants on my pooled sequencing files with CRISP which provides me with estimated allele frequency and the Hardy–Weinberg equilibrium (HWE) test is 0.0 for all SNPs. Is there anything I can do to narrow down the causative genes (SNPs) or region? or all I can do is focus on the SNPs with high allele frequency and hope for the best?

I appreciate all the insight you could provide!

Answer 1

[GenomeXP]

Hello TH Yeh,

If I understand it correctly, you are aiming to find the loci where drug resistance has evolved in three samples of flies that are resistant to the drug. Right?

In order to find such sites, you will certainly need to compare the genome of resistant flies with non-resistant flies. Even with the control, it will still be tricky to disentangle the neutral variation from the alleles of resistance.

I don't know your precise experimental design, but it seems like high allele frequency will teach you nothing since their frequency can be high just by chance, as neutral mutations can also spread in populations. As for Hardy-Weinberg equilibrium, genotypes are unknown due to the pooling so I am not sure of what you could do with this measure.

Best,

Guenole