Entering edit mode
9.7 years ago
ghbore
•
0
Dear all,
How to manual-free recognize which known transcript/isoform the alignment/mapping is, given the existence of sequencing error and alignment/mapping error? Further, judge a alignment/mapping is a novel transcript/isoform?
It seems many transcript assembly or detection softwares, like Cufflinks could provide transcript sequence (in FASTA) or exon junction (in GTF), with either no accession or their own accessions. Does anyone known any tool can handle this, or any convenient method?
Thanks in advance!
Regards,
Bore
You mean, how to automatically perform transcript mapping and how to identify novel transcripts?
Yes! Exactly, to perform transcript detection (assignment to known transcript/isoform accession number) and prediction (novel transcript/isoform), given we already know their alignment/mapping to genome. Alternatively, if it would be convinient to start from sequence and need no intermediate aligning/mapping step, it's also OK.
I have tried BLAST directly. But in some situation, I could not distinguish a bad alignment from a possible novel isoform or just sequencing error, in consideration of these sequences coming from PacBio.
What kind of data do you have? I guess it is RNA Sequencing data? If that is the case, you can follow the:
STAR -> HTSeq -> DESeq pipeline for transcript detection whereas you can use the MATS for alternative splicing analysis and you can use Trinity for de-novo assembly to identify novel transcripts
I guess you can also use cufflink, but I haven't use it for such a long time so I can't be sure.