jbh-cas
commented
1 year ago As I recall the issue is that the so-called forward primer doesn't mean that it and its barcode are necessarily at the beginning of a read. The read might have been sequenced in the reverse so the forward primer/barcode is reverse complimented at the end of the read and the reverse or ending primer/barcode would be at the beginning of the read also revcomped. So we just need the normal representation of each and make revcomped versions inside the program to look for them at either end of the read.
You might try relaxing the matching requirements depending on the quality of the reads your have. Things have improved since the program came out in regard to nanopore base quality but still suffer some issues.
Hope that gives a bit of insight.
On 10/19/2023 7:56 AM PDT sscoster @.***> wrote:
I'm using minibar to help demultiplex ONT sequences with custom indexes. I have a few questions:
a) I believe the code searches for the reverse compliment so I only need to put in indexes in the 5'-3' direction into the reference .tsv file, is this correct? b) I've noticed that when I also add reverse indexes (not compliment, just reverse of forward index, etc.), I also get reads assigned to samples. My question is, should I be adding in reverse primers to the .tsv file as well when I reverse the indexes? c) If I get reads from using the reverse indexes, is there a scenario where I should be trying all combos (reverse forward index with normal reverse index, etc.)? My thought process is that if the indexes are found reversed, the primers would also likely be reversed so there is no need for all possible combinations.
I appreciate any insights!
Thanks, Stephanie
— Reply to this email directly, view it on GitHub https://github.com/calacademy-research/minibar/issues/10, or unsubscribe https://github.com/notifications/unsubscribe-auth/AELSO4PHFNWY5PDGE46CQVDYAE5QJAVCNFSM6AAAAAA6HLKCVKVHI2DSMVQWIX3LMV43ASLTON2WKOZRHE2TENBWGQ4DCMA. You are receiving this because you are subscribed to this thread.Message ID: @.***>
sscoster
I'm using minibar to help demultiplex ONT sequences with custom indexes. I have a few questions:
a) I believe the code searches for the reverse compliment so I only need to put in indexes in the 5'-3' direction into the reference .tsv file, is this correct? b) I've noticed that when I also add reverse indexes (not compliment, just reverse of forward index, etc.), I also get reads assigned to samples. My question is, should I be adding in reverse primers to the .tsv file as well when I reverse the indexes? c) If I get reads from using the reverse indexes, is there a scenario where I should be trying all combos (reverse forward index with normal reverse index, etc.)? My thought process is that if the indexes are found reversed, the primers would also likely be reversed so there is no need for all possible combinations.
I appreciate any insights!
Thanks, Stephanie