is there a way to get splice sites from BLAST results?

Question

[wrf]

Does anyone know a program that can turn blast output into predicted transcripts with splicing sites?

I imagine some kind of program that can map blast hits (either blastn or tblastn) and then extend the hits in both directions to find optimal/ realistic splice sites.

Exonerate is unbelievably slow, and since blast already finds the best site very quickly, something that could just 'optimize' the transcripts/proteins to the genome would be great.

I find it hard to believe that something like this does not already exist.

[wrf]

To follow up, there is a specific program to do this called genBlastG, however it is long depricated (as far as I can tell); it has not been updated in years and the creator is not responding to email.

The program itself crashed for every version that I tried, so it is unlikely that this will be a viable option.

Answer 1

[Ann]

Take a look at blat - it may have an option to align protein onto genomic.

[wrf]

blat does have an option to do that, but the output is no different from BLAST, which is regions of alignment, not exons. As I understood from EvidenceModeler, raw blast output cannot be used as evidence, it needs exon boundaries so something has to convert it.