Question about logging in collapse about truvari HOT 2 CLOSED

• __filtered is the number of variants that weren't compared to other variants due to being outside of parameter thresholds (e.g. --sizemin)
• Wrote 172241 variants is how many variants are in your output vcf.
• 2103726 variants collapsed is how many variants were merged into calls that were kept. They can be found in the --collapsed-output.
• collapsed into 111970 variants is how many variants in your output VCF had >= 1 variant merged into them.

So for this example, 65% of your output variants are the representative variant from some number of merged variants. And 2103726 / 111970 or 18:1 is the ratio of how many variants collapsed into the merged representations.

I don't know if it applies to your dataset, but for variants I'm used to dealing with, 18:1 is pretty high. I suspect you've used looser thresholds (e.g. --pctsim 0). I would just caution you to look out for over merging. See our paper for details on what over merging might look like.

from truvari.

Thanks very much for your clear explanation. I've got that!
I'll check the merging according to your advice.

from truvari.