Remove adaptaters from ONT reads : An alternative to Porechop ?

Question

[Rox]

Hello everyone,

I was using Porechop (as everyone I guess) in order to remove adaptaters from ONT reads. This tools is now unsuported (Removing adapters from Oxford nano pore reads , https://github.com/rrwick/Porechop), and with rapid changes in ONT softwares, I'm really afraid it will soon be impossible to use this tools anymore, and I'm looking for alternativ tools.

I found the new tools of Ryan Wick Deepbiner ( https://github.com/rrwick/Deepbinner ), which is presented by him as the sucessor of Porechop, but it is unclear if the tools only demultiplexes reads or also removes the adaptaters from them. For example, I feel like it's impossible to use on non barcoded runs.

Does anyone knows more about that subject ? I've been looking for some ideas on Nanopore Community, but so far my question remains.

Cheers,

Rox

Answer 1

[WouterDeCoster]

Porechop may be no longer maintained, but it still works. I don't think there is any reason to quickly start looking for alternatives. Based on what I read on the nanopore community forum the adapter removal step would be added to Guppy, the basecaller which is to replace albacore in the "near future". So they (finally) start supporting this themselves.

[Rox]

"Near future" ;) Alright, I see,

Thanks for the insight then. I'll keep play with Deepbiner tho, because it claims the result contain less than the 30% unclassified reads... Did you ever tried that tool ?

[WouterDeCoster]

To be honest I don't use barcodes all that often - just one genome per PromethION flow cell ;-)

[Rox]

I see ;) Bacterial genome with 15000X for the win o/

[WouterDeCoster]

When do we get these damn flongles :o

[Rox]

The hype is real o_o

Answer 2

[Michael]

While it's maybe all fine and good that Porechop is still running after 6 years, the worst problem may be that the adapter sequences are not up-to-date. Then the best alternative to porechop is possibly your own github fork of porechop, modify adapters.py with sequences from the current protocols. My fork is here btw. I just updated the adapters, not the barcode sequences, and didn't care about redundant adapter sequences.

Also, adapter trimming might not be relevant at all anymore depending on base caller and settings so the other alternative might be to not run adapter trimming at all.