search

DaehwanKimLab / hisat2

Graph-based alignment (Hierarchical Graph FM index)
GNU General Public License v3.0
464 stars 113 forks source link

Methylation output of hisat-3n #331

Open zhangzhiyangcs opened 2 years ago

zhangzhiyangcs commented 2 years ago

Hi, I used to utilize the bismap software and Its equipped scripts such as methratio.py which give me CG, CHG, CHH sites. But this scripts is not worked on hisat-3n alignment files. Is hisat-3n equipped with function? I try the hisat-3n-table command and the result need to further processing. Do you any suggestions? Best wishes

imzhangyun commented 2 years ago

Hello Zhiyang,

Thank you using HISAT-3N. Unfortunately, current hisat-3n-table does not classify whether the site is CG/CHG/CHH. You can create a simple script to label the bases in 3n-table file: The base with strand '+' is C, the base with strand '-' is G.

Leo

zhangzhiyangcs commented 2 years ago

Hi Yun, Thank your software which get much better mapping rate than bsmap and bismark in autopolyploid plant genome (4X) which achieve 95% rate vs 17% rate. I don't understand the reason. Is your algorithm more suitable for autopolyploid plant genome? In addition, Is there easy way to export CHG or CHH site just like --CG-only? And should I do the futher Binomial test to checkout the really methylation site?

imzhangyun commented 2 years ago

Hello Zhiyang,

What the the percentage of your reads unique aligned to the reference? HISAT-3N can report both unique and multiple aligned reads, in contrast, Bismark can only report unique aligned reads.

Leo