tetraploid de novo species

[Silvia Bettencourt]

Hi everyone, 

Does anyone knows if stacks2 is set to call SNPS and variants on tetraploid species? I am running the optimization parameters step, and the R80 rule does not stop giving me increased kept loci, even at M/n of 8 and m=2.  Does anyone knows any "trick" to make this work? I know the Stacks parameter optimization is not a make or break rule but I would like to, at least, have some indications of a more favorable parameters set. This is a non-model species running DeNovo.

[Catchen, Julian]

Hi Sylvia,

 

The SNP calling model is designed around diploids. However, it is not really unusual to see R80 continue to add loci at an M/n of 8 and higher (however, we strongly recommend an m=3 and not to vary this parameter). You are welcome to post your optimization numbers, but it is really a question of how variable the results are. That is, if you are adding small numbers of loci at M/n of 9 vs 8, or 10 vs 9, that typically means they are fairly similar and you will get similar results from any parameter choice. If, on the other hand, you have really different results at those different points, you may need to keep searching at higher values of M/n or consider other things that may be going on with your data.

[Silvia Bettencourt]

Hello, thank you for your reply and time. I´ll redo the optimization with m=3 as you advised. Regarding my current optimization values set for m=2, M/n=1 produced 3296 of kept loci, M/n=2 added 204 loci, M/n=3 added 106 loci, M/n=4 added 285 loci, M/n= 5 added 127 loci, M/n= 6 added 162, M/n= 7 added 250 loci and M/n=8 added 121 loci. As soon as I have the m=3 optimization completed I´ll post the results.

[Silvia Bettencourt]

hello sir, god evening.

just as a follow-up. i finished running my samples from M/n from 4 to 12 and the kept loci does not drop to negative. the additions are as follows: +57; +68; +190; +103; +41; +53; +28; +77 with a start point of kept loci at M/n =4. At this point i would appreciate suggestions. I´m gonna stop and revise everything from scratch as it seems something is biasing my results. i find M/n =12 a bit high although this is a tetraploid species.

Any suggestions? this regards a tetraploid species, endemic, with discrete distribution, paired-end and single digested RAD-sequenced by ILLUMINA. 24 samples were sequenced, and the same 24 samples were used in the optimization. The original output per sample goes from 85Mb to 4.5Gb.

best regards

Silvia