SEQanswers (
-   Ion Torrent (
-   -   Bismark for Ion Torrent (

bioWizz 08-18-2014 09:00 AM

Bismark for Ion Torrent
I am trying to use Bismark on targeted bisulphite sequencing done on PGM. Average read length is about ~102 maximum being ~190.
Bismark uses bowtie2 and running it with varying parameters gives me aa lignment score of 8 - 10 % which is very low. When I use TMAP on torrent suite to map to converted reference sequence, I get decent alignment percentage.
Any suggestions on bowtie2 parameters for Ion Torrent variable read length data? Suggestion on analysis of targeted bisulphite sequencing with PGM output will also do.
Thanks :)

frozenlyse 08-18-2014 11:09 PM

I wouldnt use bismark on ion torrent data as it is going to have a lot of indels due to homopolymers, and bismark doesnt run bowtie2 in local alignment mode - I would try as it uses bwa for the alignment - however it may require paired end reads!

bioWizz 08-18-2014 11:41 PM

As you mentioned I had already tried running bowtie2 with local alignment mode, but it did not work.
Just looked into BWA-meth, it does say it supports single end reads. Looks like the tools I was looking for as it employs a similar method as Bismark. Will try to use it and update.
Thank you :)

dpryan 08-19-2014 12:37 AM

It can be a pain to tweak bowtie2's local alignment setting to get decent results (that's one of the reasons I have Bison default to end-to-end alignment). BWA-meth's defaults are pretty reasonable and might give decent enough results, though as frozenlyse mentioned, ion torrent data is often riddled with InDels. Brent's pretty good with timely feedback, so you might just drop him an email if you don't get decent results.

BTW, you can also try sending me a small test dataset and I can have a look with Bison and see if I can send you some reasonable settings, though I expect BWA-meth will work.

IonTom 08-19-2014 02:58 AM

I don know if the original BWA works Ion Torrent data. I am currently using BWA-Mem and it works fine. Is there a methylation mapper based on Mem ?

There is also NextGenmap which works well for Ion Torrent data and also includes
a methylation mapper.

dpryan 08-19-2014 03:28 AM

@IonTom: See the aforementioned bwa-meth package from Brent Pederson.

vz33 11-07-2014 01:41 PM

@bioWizz and @IonTom, which kind of bisulphite sequencing do you did by PGM? RRBS, TAB-Seq or others? Could you share a protocol to me? Thank you.

All times are GMT -8. The time now is 03:57 AM.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2019, vBulletin Solutions, Inc.