Hello dear developers, I was checking the zymo BS protocol and I found a little discrepancy with the Methylseq adapter trimming parameters: https://files.zymoresearch.com/protocols/d5465-zymo_seq_wgbs_library_kit.pdf the protocol reports "the first 15 nucleotides of Read 1 and Read 2 should be removed after adapter and quality trimming has been completed" however the

--zymo

parameter is equivalent "`--clip_r1 10`

--clip_r2 10

--three_prime_clip_r1 10

`--three_prime_clip_r2 10`". I was wondering why

this preset is probably there for a very long time, and my guess is that often, removing 10bp from each end of the reads is sufficient to remove all or most of the apparent biases of the protocol. Instead of using the preset you could of course set all 4 trimming options manually (e.g. to remove 15bp each) (don’t forget to set

--non_directional

, too). You should in any case check the methylation bias (M-bias) plot once the run is complete. This is the best way to judge whether there are any residual biased positions left in the reads. If you find that there are some leftover biases (consistently), we could change the default settings to be even more conservative.

Thank you for the clarification Felix. Can you explain why you suggest to trim from both ends instead of the 5' end only? I ask because that's not mentioned in the zymo protocol

In my experience, bisulfite fragments a are typically fairly short, probably somewhere between 70 and 200bp (you can check this in your case). Let’s assume your fragment is 110bp long, and your sequencing is 2x150bp, each of your reads would have the bias at the start (say 10bp), but then still be long enough to read through the bias on the other end, and then into the adapter. Something like so::

-----------------------------------------                                    sequenced fragment length  
--------------------------------------------------------------------------------------> read length
BBBBBBBBBB FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF BBBBBBBBBB AAAAAAAAAAAAAAAAAAAAAAA read
BBBBBBBBBB FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF BBBBBBBBBB                         adapter trimmed read
           FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF                                    adapter + bias trimmed read

With

B

being bias,

F

being fragment of interest, and

A

being read-through adapter contamination.

Thanks a lot! Very informative 👍 headbangingparrot

headbangingparrot is possibly the most excited reaction to one of my trimming posts I have ever seen 🙂