look for "Pool-GWAS" in google scholar, there are lots of examples. This is indeed normally done for inbreed lines so I assume you will have to have both very large numbers of individuals in your pools and very strongly selected traits for this to work

Hi Pritam,

Genenetwork webqtl won't work for me, the others discussions linked to your answer are helpful though.

Thank you.

Hernan Morales ,

Thanks for pointing in the right direction. Do you have any personal experience with Popoolation 2?

Arijit Ghosh I do, but not with quantitative traits. Happy to help with generalities if you have any questions. Cheers

Hi Arijit, it seems like Popoolation2 should work for your experiment, however I would say it is easier when you have created a mapping population. I wonder how did you get the 2 selected population? are they coming from a single population?

Hi Lucio Navarro , Yes, I am working out how Popoolation2 can hep me, seems promising. About mapping populations, we have not created any, our selected populations are also large outbreeding ones. And between the 2 populations I have mentioned, one is control and the other is the selected one (both are large outbreeding populations with almost equal numbers of individuals).

Well, Popoolation2 should work for. You will need to be stringent during SNP calling. As Hernan Morales says, large number of individuals in the pools and strongly selected traits will help. The best.

Hernan Morales and Lucio Navarro , do you have any idea how to go back to VCFs from CMH test results? Equivalent SNPs in all replicates will be in CMH results whereas I can not make a synced VCF for all replicates.

Hello Arijit Ghosh , I personally do not have experience with CMH, but looking at this tutorial https://sourceforge.net/p/popoolation2/wiki/Tutorial/ I don't understand how you can have CMH results without having the sync file (sync files are not VCFs, though). You need sync files to do the CMH test, so I'm confused why you need to go back to sync files?

Hernan Morales , I made sync files and that is how I could run the CMH test. My concern is when you run the CMH test, it gives you consolidated report for ALL replicates, not replicate-wise results, and that makes sense as the test is designed to find out the significant difference in allele frequencies common to all replicates of the control and experimental populations. Now, for further downstream analyses, I need to look at the SNP identities, and their putative effects on the gene or gene product (like SNPeff etc). I can only do that if I have the SNP identities for the locations of each replicate, that information is lost during CMH test.

Let me give you an example:

2L 5189 N 1:33:127:0:0:0 2:18:165:1:0:0 4:38:156:1:0:0 1:15:98:0:0:0 1:30:126:0:0:0 0:24:113:1:0:0 1:23:101:1:0:0 0:0:104:0:0:0 4.301062e-05

This is the result from a CMH run between two populations having 4 replicates each. Now, how do I go back to the SNP identities? And predict their putative effects?

Hello Arijit Ghosh the sync file are simply read counts. In your example you are testing your SNP in chromosome 2L position 5189. So you can either go back to your annotation and see where this SNP is or if you want to infer anything related to its frequency you can estimate the pool allelic frequencies with the script snp-frequency-diff.pl

This is an example output showing the allele states and the read counts / total reads, from which you can obtain the frequency. I guess this is what you need to see if there was a replacement, etc at a given position.

``` OUTPUT _rc

##chr pos rc allele_count allele_states deletion_sum snp_type major_alleles(maa) minor_alleles(mia) maa_1 maa_2 maa_3 maa_4 maa_5

mia_1 mia_2 mia_3 mia_4 mia_5

gi|42519920|ref|NC_002978.6 3530 A 2 A/G 0 pop AAAAA GGNNG 54/55 76/78 35/35 27/27 17/18 1/55 2/78 0/35 0/27 1/18

gi|42519920|ref|NC_002978.6 5450 T 2 T/A 0 pop TTTTT ANANN 51/54 96/96 9/10 33/33 23/23 3/54 0/96 1/10 0/33 0/23

```

Hernan Morales , thank you so much. I completely forgot this info will be in the rc and also some relevant information in the pwc. People like you make researchgate an awesome experience.