Trim TSO on read end

[Zepeng-Mu]

Hello,
I'm very new to ONT data analysis and flexiplex has been really helpful to us!
I wonder if it's possible to trim TSO sequence on the other end of read, so that it is not used for aligning (although minimap2 should handle it very well). Basically, the read looks like below:

5'---[flank]-[CB]-[UMI]-[polyT]---[transcript]---[TSO]---3'

Currently I can extract CB and UMI without problem, although the TSO remains in the reads. I'm using cutadapt to remove the TSO for now. It would be great if flexiplex has a way to remove it.

Thanks!

[nadiadavidson]

Hi,

I'm glad flexiplex has been helpful to you.

We don't have a way to cut off the TSO at the same time as extracting the barcode and UMI. However, you could run flexiplex separately to remove the TSOs (either before or after the barcode demultiplexing). Something like:
flexiplex -x [reverse compliment of TSO] -d grep -r true -f [maximum errors in TSO seq]
could work. Here the -x [reverse compliment of TSO] -d grep tells flexiplex to search for the reverse compliment of the TSO, and the -r true tells it to trim. The sequence before the search string is always trimmed, which is why you may need to search for the reverse compliment (to switch the transcript to the right side).

If you wanted to run the trimming and barcode demultiplexing in one go you could do something like:
flexiplex -x [reverse compliment of TSO] -d grep -r true reads.fastq | flexiplex -d 10x3v3 -k [your barcode list]... > final_reads.fastq

A couple of notes here: 1) I haven't tested this code at all , so it might need some tweaking! 2. Any reads without a TSO won't be reported at all, so if you expect some without, and want to keep them, you might be better using just cutadapt or another tool.

Please let us know if you have any questions.

Cheers,
Nadia.

[Zepeng-Mu]

Hi,
Thanks, this is very helpful!
I thought about -d -r but did not realize I can use reverse compliment of TSO.
I guess now I'm not sure if I should remove TSO at all. Looking at the 10x presets in flexiplex there is no TSO matching. I wonder what is the rationale for doing so, is it because many useful reads may not have TSO, and for those do have TSO minimap can easily softclip those bases?

Best,
Zepeng