Subsetting gt_donor.vcf

[LuisHeinzlmeier]

Is there a reason we want to filter gt_donor.vcf and not gt_cell.vcf? Subsetting gt_donor.vcf only works when the method vireo is called.
What can the subset of gt_donor.vcf be used for?

[Zethson]

I have no clue. We'd have to ask Fabiola

[LuisHeinzlmeier]

During donor matching, the pipeline can also determine two subsets of informative variants (default: subset as described in paper; vireo: subset by Vireo). After that gt_donor.vcf was subsetted with these variants (bcftools filter ${prefix}_sorted.vcf.gz -R $subset_variants -Oz -o ${prefix}_filtered.vcf.gz). Finding these variants and subsetting the vcf was only possible if vireo was the best method in donor matching.

gt_donor.vcf is produced by vireo and shows the genotype for each donor.

#CHROM  POS     ID      REF     ALT     QUAL    FILTER  INFO    FORMAT  jurkat  293T_RTG
21      30408670        .       A       T       .       PASS    AD=92;DP=197;OTH=0      GT:AD:DP:PL     1/0:42:93:100,0,100     1/0:50:103:100,0,100
21      34621124        .       G       A       .       PASS    AD=1;DP=23;OTH=0        GT:AD:DP:PL     1/0:1:6:55,0,100        0/0:0:17:0,69,100
21      34669381        .       G       A       .       PASS    AD=22;DP=34;OTH=0       GT:AD:DP:PL     1/0:10:16:100,0,100     1/0:12:18:100,0,100

gt_cell.vcf is produced by cellsnp-lite and shows the genotype for each cell.

#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT AAACGCTGAAGATG-1 AAAGATCTGCGTTA-1 AAAGCCTGTCACGA-1 AAATA>
21 30408670 . A T . PASS AD=92;DP=197;OTH=0 GT:AD:DP:OTH:PL:ALL .:.:.:.:.:. 0/0:0:1:0:0,3,19:1,0,0,0,0>
21 34621124 . G A . PASS AD=1;DP=23;OTH=0 GT:AD:DP:OTH:PL:ALL .:.:.:.:.:. .:.:.:.:.:. .:.:.:.:.:. .:.:>
21 34669381 . G A . PASS AD=22;DP=34;OTH=0 GT:AD:DP:OTH:PL:ALL .:.:.:.:.:. .:.:.:.:.:. .:.:.:.:.:. .:.>

In the current implementation of the pipeline, finding variants (at least for default: subset as described in paper) is always possible in rescue mode because cellsnp-lite is always called. Not only for vireo. However, subsetting gt_donor.vcf, of course, only works when vireo was running. So I thought: Is there a reason why we only want to subset gt_donor.vcf? We could also subset gt_cell.vcf ... As I don't know what the subsetted vcf can be used for, I created this issue. In the paper you mentioned "an additional genetic demultiplexing". But for this we need gt_cell.vcf as input and not gt_donor.vcf, right?

The relevant part in the paper
The relevant part in docs

[LuisHeinzlmeier]

We could also subset the input VCF and rename --subset_gt_donors to --subset_vcf.