Brent Pedersen

I think 50 would be the very lower limit where it could be expected to be reliable, but even that is a bit small.

you can see in the readme the evaluations for >300 and >1000 but I think the accuracy should still be high above about 100 or so.

Hi, `encode` takes the VCF as the final argument, not the first. So you need to flip the order of the .zip and .vcf I do not have hg19 gnomad....

Hi, sorry for the delay, if you're still having this problem, Can you show the header of the bam file, specifically the SN tag for chr1? You could also show...

Thanks for reporting. Using the lower --proportion-repeat does indeed uncover some problems. I can recreate and pushed a fix. We'll need to evaluate the change in performance. I also have...

BTW, I'm interested to hear what @hdashnow thinks, but my experience is that lowering proportion_repeat is usually not fruitful.

I would say to make sure that `methylation_cfDNA/resources/bwameth/hg19.fa` (a relative path) actually exists and that all the .c2t and index files exist in the directory.

that looks ok. my guess it that there's something odd with how you're specifying the path. maybe the part of the path you have elided has spaces or is incorrect...

networked file system

the argument to --reference should be the original .fasta. so e.g. ``` --reference hg38.fa ```