Using ARPIP to impute/reconstruct missing sequences from an MSA? #4
Description
Dear Dr. Gholam-Hossein Jowkar,
How this thread finds you well, I tried to reach you at your academic email address ([email protected]) but the mail delivery system returned an error, so I decided to reach you here.
I'm writing to ask for some guidance/heads up regarding the functionality of "ARPIP" and whether it can be leveraged to "impute" systematically (aka not sparse nor random) missing partitions from a set of aligned sequences... I only found a tool that explicitly says it can: ForeSeqs (D. Darriba & A. Stamatakis). However, after some testing I realized that it can't handle indels, and it always predicts a nucleotide for all missing/unknown sites even if such site probably should correspond to a gap (example attached).
This behaviour is not what I was looking for, so I came across your software, and it seemed like it could have worked, but then it appeared like ARPIP does not work with missing sites(? which was a little sad...
However, from the "Features and project structure" section:
| "Please note that unknown character is not supported in this version of ARPIP."
This leads me to believe that perhaps you may consider adding in the future something to deal with ambiguous/unknown characters such as imputing them in the MSA using something like information from the ancestral sequence, branch lengths, etc.? or is there a way to tweak the code in order to do so?
Just want to know if you think about adding new features like that to ARPIP, so I can manage my hopes haha... outside of ForeSeqs (which doesn't handle gaps at all and doesn't seem maintained at all), and the (in my opinion) poorly documented BAli-Phy, I don't think there are any other tools that may be able to perform this task. If you happen to know of another one I'd appreciate your insight.
Thank you for your attention.
Sincerely,
Alan Vladimir
Undegraduate Program on Genomic Sciences, UNAM