lh3 / psmc

Implementation of the Pairwise Sequentially Markovian Coalescent (PSMC) model
Other
156 stars 60 forks source link

recommended min and max for vcfutils.pl step #17

Open sjfleck opened 4 years ago

sjfleck commented 4 years ago

Thank you for this interesting and useful program. I have a question about this step:

samtools mpileup -C 50 -uf ref.fa aln.bam | bcftools view -c - | vcfutils.pl vcf2fq -d 10 -D 100 | gzip > diploid.fq.gz

In the README file, it recommends that -d should be set to 1/3 of the average read depth and -D should be set to 2x the average read depth. I successfully completed PSMC using default values for these parameters and again with the recommended values based on my sample's average read depth of 17x (many of my samples are around this value). The resulting plots were very different despite only changing the min and max read depth.

I shared a picture below with. "SFG00015-A.elasticus" has default values (-d 3 & -D 100000) and "SFG00015-A.elasticus-test" has the recommended values (-d 5 and -D 34) for the same sample. I'm wondering which curve I should trust. Is my coverage too low to use the recommend values?

Any insight into this would be greatly appreciated. Thank you

ARTOCARPUS_PSMC_t15.pdf