search

nanoporetech / dorado

Oxford Nanopore's Basecaller
https://nanoporetech.com/
Other
452 stars 53 forks source link

Dorado Alignment circular sequences #712

Closed Macdot3 closed 3 months ago

Macdot3 commented 3 months ago

Hello everyone, I wanted to ask a question about alignment with Dorado. I'm analyzing circular mtDNA sequences and I was wondering if it's possible for Dorado to recognize such sequences as circular without any issues of read shifting at the end and beginning once aligned, or even worse, loss of information. It would be very helpful for us working with these sequences. Alternatively, is there a way to overcome this? Thank you very much for any advice.

beppybeeps commented 3 months ago

Hi, I'm not aware of any aligner that handles circular sequences well but I work around the issue by copying the entire sequence and appending it to the end. This way a full length read can start from any point and always has somewhere to map continuously. It does make downstream analysis a bit more complicated though so I'd be happy to hear about a better option.

tijyojwad commented 3 months ago

Hi @Macdot3 - dorado aligner calls into minimap2, so dorado aligner can only do what minimap2 is able to do. It might be a good question for the minimap2 repo. If it's possible to handle circular genomes with minimap2, then we can do it with Dorado aligner too. And if there are options missing in dorado to handle that, we can look into adding those.