Adapter trimming about gembs

[heathsc]

Not normally, unless you are expecting to see a large fraction of DNA fragments that are less than 1 read length. In general the GEM3 mapper is quite good at mapping even with quite high levels of adaptor contamination. Simon

…

On 28 Feb 2018, at 13:22, berguner ***@***.***> wrote: Hello, Is adapter trimming required or recommended for WGBS analysis using gemBS? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#12>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADHPd0BBnwU2zavaGR_cS7WWFyGgoSo_ks5tZUURgaJpZM4SWkOy>.

[berguner]

Hello Simon,

We have a library which was prepared for 75PE sequencing so the insert size was short. We also sequenced this library with 150PE on NovaSeq to test the instrument. As expected, there is a lot of adapter contamination in the 150PE data. Although gem3-mapper aligns most of the reads regardless of the adapter contamination, the rate of unique reads decreases considerably. For the same sample the average unique rate increases from 44% to 54% after adapter+3'quality trimming whereas it is 74% in the 75PE data. The rate for 75PE is much higher because GEM3 mapper gives 0 mapping quality to 100% overlapping pairs. Moreover, the adapters are aligned with gaps and mismatches rather than soft clipping which might -IMHO- increase the FDR of variant/bisulfite calling. Please have a look at the IGV images below where you can see the difference between 75PE, adapter trimmed 150PE and non-trimmed 150PE data (from top to bottom respectively).
https://drive.google.com/file/d/1oG7Y1wRxKEjarDN5HC1qg_MkuIxpqztQ/view?usp=sharing
https://drive.google.com/file/d/1yrXTMHkK1xZRz1BWcBe03Qit7BeBr-Fa/view?usp=sharing

Best,
Bekir

[heathsc]

Hi Bekir, Thanks for the report. The libraries that we see mostly have many pairs where the two reads overlap, but very few where the insert size is less than a single read-length. In this case GEM3 will not set bit 2 in the SAM flags (indicating ‘proper’ alignment), and the read pair will be discarded by bs_call. I could add a flag to the caller to not discard such read pairs and that should (hopefully) resolve many of the problems that you are seeing. Simon

…

On 6 Mar 2018, at 10:17, berguner ***@***.***> wrote: Hello Simon, We have a library which was prepared for 75PE sequencing so the insert size was short. We also sequenced this library with 150PE on NovaSeq to test the instrument. As expected, there is a lot of adapter contamination in the 150PE data. Although gem3-mapper aligns most of the reads regardless of the adapter contamination, the rate of unique reads decreases considerably. For the same sample the average unique rate increases from 44% to 54% after adapter+3'quality trimming whereas it is 74% in the 75PE data. The rate for 75PE is much higher because GEM3 mapper gives 0 mapping quality to 100% overlapping pairs. Moreover, the adapters are aligned with gaps and mismatches rather than soft clipping which might -IMHO- increase the FDR of variant/bisulfite calling. Please have a look at the IGV images below where you can see the difference between 75PE, adapter trimmed 150PE and non-trimmed 150PE data (from top to bottom respectively). https://drive.google.com/file/d/1oG7Y1wRxKEjarDN5HC1qg_MkuIxpqztQ/view?usp=sharing <https://drive.google.com/file/d/1oG7Y1wRxKEjarDN5HC1qg_MkuIxpqztQ/view?usp=sharing> https://drive.google.com/file/d/1yrXTMHkK1xZRz1BWcBe03Qit7BeBr-Fa/view?usp=sharing <https://drive.google.com/file/d/1yrXTMHkK1xZRz1BWcBe03Qit7BeBr-Fa/view?usp=sharing> Best, Bekir — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#12 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADHPd4pH5Mj_ynULyTAXJV9ClxxtDIFaks5tblQXgaJpZM4SWkOy>.

[berguner]

Thank you Simon, it would be great if you can implement this soon.

Best, Bekir

[berguner]

Hello Simon,

I could add a flag to the caller to not discard such read pairs (100% overlapping)

Did you have a chance to implement this feature?

Best, Bekir

[heathsc]

Hi Bekir,

In principle the new version of gemBS (3.0.0) should have this feature. The --keep-unmatched option to gemBS call will not discard read pairs that are not pairing correctly.

Simon