bayenv2: "initial epsilon zero or 1 locus 6912"

Question

[biosol]

Hi everyone,

currently I'm running bayenv2 on my data. First,I've converted my vcffile into bayenvformat but then, when I try to generate the covariance matrix, I get the message initial epsilon zero or 1 locus 6912. I've read that the epsilon value in bayenv2is related to the allele frequency at a specific locus for the current generation, but don't actually understand what this means and I don't know how to solve this problem.

Any help will be appreciated! Thank you very much in advanced :)

Answer 1

[katuckele]

Hi Sonia,

I had this error message too. I checked my SNP file and sure enough, the locus referenced by the error message was fixed in all populations, e.g.:

0 0 0 0 0 0

20 20 20 20 20 20

I took out this locus and Bayenv runs without error.

Hope this helps, Katie

[biosol]

Hi Katie, thank you very much for your answer! I asked the developpers of the bayenv2 and they told me that the problem was that the file contained SNPs with fixed allele frequencies (the same allele in all individuals, either 0 or 1). So I filtered my data by MAF 1% and the problems was solved. I think this matches as well with your case, your allele freqs were fixed for that SNP!

[QPaps04]

Just to add that I also recieved this error when I had a typos in the sample names of the popualtion file I used to convert vcf to BayEnv using PGDSpider. It looked as though it converted correclty but then when I ran BayEnv this error occured. Just in case people still get this error after looking for fixed sites!