Using bed file to identify callable sites

[juanitagutierrez]

Hi :)

I would like to understand how to use the bed file produced by gimbleprep to reproduce the VCF generated by gimble after this first step. It seems like the bed file lists all loci, e.g.,

chr1 0 137 10 sample_1,sample2,sample3,...,sample10
chr1 137 139 11 sample_1,sample2,sample3,...,sample11
chr1 139 169 10 sample_1,sample2,sample3,...,sample10
chr1 169 177 11 sample_1,sample2,sample3,...,sample11
chr1 177 180 12 sample_1,sample2,sample3,...,sample12

Should one pre-filter the bed file to keep only sites supported by more than N samples?
Thanks in advance for your help.

Best,

[GertjanBisschop]

I might be missing something here but at which stage does gimble produce a vcf? If you mean gimble before we split off gimbleprep, then the answer is that gimbleprep + gimble = old gimble. So the output of the entire workflow should be the same.

[juanitagutierrez]

Hi again,

Thanks for your prompt answer.
I guess my question wasn't clear: after running gimbleprep, one gets a set of files for downstream steps, including: *genomefile, *coverage_summary.csv, *bed and a *vcf.gz (plus its index). I understood (maybe wrongly) that by using the *bed file, one could see which loci were kept after processing the vcf with gimbleprep, and that it should be possible to use it if one wanted to, let's say, filter another vcf to keep loci identified as callable by gimbleprep. This is why I was curious to know if that *bed callable sites should be filtered in some way to do this.
I hope I am explaining myself better this time!

Best,

[GertjanBisschop]

The *.bed file indeed contains a list of genome intervals and the samples for which this particular genome interval can be used for further analysis. So you are right that you could filter another .vcf using this .bed file as long as the sample labels listed in the .bed file are also present. I am not sure though whether why you would want to do this. Wouldn't the .vcf for another dataset have different coverage values for each of the intervals and samples?