search

KChen-lab / Monopogen

SNV calling from single cell sequencing
GNU General Public License v3.0
68 stars 16 forks source link

Different somatic calls across Monopogen v.1.0 and v1.6.0 #58

Open ypershad opened 2 months ago

ypershad commented 2 months ago

We are following the example here and noticed the following overlap with the example data between somatic SNV calls made from scRNAseq with v1.0 and v1.6.

image

Are you surprised that ~25% of the variants differ between versions? Do you think that the v1.6 calls are more trustworthy?

Thanks for any input that you have and for building this wonderful tool.

jinzhuangdou commented 2 months ago

Thanks for you posting such comparison. Your testing is based on our chr20 MAESTER dataset or yourself? I will look at a bit and feedback to you soon.

ypershad commented 2 months ago

Based on your chr20 MAESTER dataset. Thank you!

jan-engelmann commented 1 month ago

Hey, do you have an update here? 🙏

jinzhuangdou commented 4 weeks ago

Hi, could you try the new version (updated in Jun 23,2024)? Only running the LDrefinement step is okay. The current version v1.6.0 imposes more filtering than v1.0.1. In v1.0.1, I noticed there are some loci with multi-alleles. This leads to the inconsistence on sequencing depth counting. The new version has fixed this issue.

jinzhuangdou commented 4 weeks ago

Another is you can check whether the discordance loci between two versions are both low sequencing depth. The updated version is the optimal one.