search

FelixKrueger / Bismark

A tool to map bisulfite converted sequence reads and determine cytosine methylation states
http://felixkrueger.github.io/Bismark/
GNU General Public License v3.0
386 stars 101 forks source link

Convertion rate #357

Closed RubioB closed 4 years ago

RubioB commented 4 years ago

Hi Felix,

I carried out the verifications of the convertion rates using the phage lambda and the chloroplast genome. With lambda phage I have a convertion rate close to 100% however with the chloroplastic genome I am at 98.8 %. What does this mean for the rest of my analyzes ? Can I still use my data ? How can/should I take this into account ?

Thanks in advance !

Bernadette

FelixKrueger commented 4 years ago

I personally would just acknowledge these rates, keep them in mind - and move on. I suppose it could mean that in 1 out of 100-1000 unmethylated positions you might observe one position methylated as a consequence of non-conversion rather than being actually methylated.

For real genomic positions (as opposed to an unmethylated spike-in control) you do however also have true CpG and non-CG methylation that are indiscriminable from non-conversion events. And furthermore, there are some positions that appear to be chemically protected from conversion, see e.g. here: Evidence Suggesting Absence of Mitochondrial DNA Methylation.

So bottom line: your level of conversion looks fine, but if all your major changes are in the region of up to 1% methylation diffference you need to be aware that this is potentially also the region of bisulfite non-conversion error.s