identify duplicated transcripts from blast report

[illinu]

I am working with a transcriptome assembled de novo. After a blastx analysis I kept the besthits (one hit per isoform) from which I isolated those annotations that are present more than once belonging to isoforms that are either duplicates or non-full-length (assembled in more than one contig).

How can I discern which isoforms are duplicates or multiassembled based on the start-end position of the annotation?

Any tools or scripts out there?

Thanks
Il.

[GenoMax]

What format is your data in (GTF)?

[illinu]

The blast report is a tab separated file, the transcritps in fasta, but I am also producing a GTF3 file with another annotator

[GenoMax]

Depending on what your files look like you may be able use some combination of unix utilities (grep/sort/uniq/awk etc) but otherwise this may need custom code.

Have you tried doing some basic sorting in excel on the GTF file?

[illinu]

Yes, when I sort by the subject hit I clearly see when two or more isoforms are duplicated or not (when the subject start-end position overlaps between the two isoforms) or multiassembled (when each isoform hits a different part of the subject), but there are 40k isoforms and I want to automate the analysis.

Using grep/sort/uniq/awk would be great. I'v done all the previous filtering with those tools but I don't see how I could use them with what I want to do now.

Il.