search

browning-lab / flare

The flare program performs local ancestry inference
31 stars 7 forks source link

Losing a High Percent of Markers #14

Closed Atvicenti closed 5 months ago

Atvicenti commented 5 months ago

Hello!

I've noticed a significant drop in markers from my input VCF file containing genotype data and the FLARE output VCF.

The min-maf/min-mac is set to 0, and there appears to be a significant number of markers in my reference.

Is the drop in markers expected? I am not sure if there is internal FLARE filtering, or if there may be something wrong with our pipeline.

Sorry if the question is pretty vague, but would appreciate any guidance!

browning-lab commented 5 months ago

Markers must be present in both the reference VCF file and the input VCF file in order to be analyzed. Markers are considered the same if the VCF records in the reference and input VCF files have identical CHROM, POS, REF, and ALT fields. Markers that are found in only in one of these two VCF files are not included in the flare output file.

Atvicenti commented 5 months ago

This makes perfect sense.

Upon re-evaluation, I found there was significantly less overlap than expected between the reference and the input.