Write out transcripts with corrected splice junctions

[msto]

Hi,

I've attempted to use TranscriptClean to correct noncanonical splice junctions, but the sequences and cigar strings in the output remain the same as in the input. After stepping through the code, I believe it is because the call to cleanNoncanonical does not update the transcript objects in the noncanTranscripts dictionary before they are written out.

https://github.com/dewyman/TranscriptClean/blob/71e7572d91bff5268d265824aee531b4292a9802/TranscriptClean.py#L163-L171

Note that the cleanNoncanonical function constructs a copy of the transcript object and updates its attributes with the corrected sequence and cigar string, but the original transcript in the dictionary is not replaced with the corrected transcript. (The Transcript2.addCorrected_NC_SJ function only logs the number of transcripts with a corrected splice junction.)

https://github.com/dewyman/TranscriptClean/blob/71e7572d91bff5268d265824aee531b4292a9802/TranscriptClean.py#L717-L729

I've changed the cleanNoncanonical function to replace each transcript in the dictionary after it's corrected.

Matt

[dewyman]

Hi Matt, thank you so much for catching this. I have verified in some of my own data that you are correct. The bug must have been introduced when I was trying to accomodate some of the edge cases people reported, because I know that this was not a problem in the version we published in our paper. I'll accept your pull request and release a patched TranscriptClean version immediately.

[msto]

Hi Dana, of course! Thanks so much for the quick merge.