Aligning genomic reads to transcriptome assembly

[ffsinga]

I have a Trinity assembly for a eukaryotic transcriptome (default parameters) and Illumina PE genomic shotgun reads (unassembled). I want to verify that the results of my assembly exist in my genome.

I have been blasting particular transcripts against the genomic reads, pulling out the reads and aligning them back to the transcript to look at coverage and verify whether there is assembly error. This is taking a long time and is not general enough for me to apply to 3 other transcriptome-assembly-genome-reads comparisons I would like to do.

Suggestions for methods to try?

I was thinking of using Exonerate to align genome reads back to transcriptome, but I'm not sure if that will take into account intronic sequences that are present in my genomic reads but not in the "reference" (the transcriptome assembly).

[Jan_R]

I think I understand what you want to do. I cannot come up with a working solution but maybe TopHat can help you? http://tophat.cbcb.umd.edu/manual.html
I have never used this program but the description sounds as if it could fit your issue.

[Wallysb01]

Why not just align the transcriptome to the genome using exonerate?

Trinity also has a reference guided transcriptome assembly.

What is it you are looking for when aligning the genome to the transcriptome?

[Jan_R]

...
I was thinking of using Exonerate to align genome reads back to transcriptome, but I'm not sure if that will take into account intronic sequences that are present in my genomic reads but not in the "reference" (the transcriptome assembly).

I think the idea is to find artifacts in the assembly.
Any aligner that is able to open gaps should fit for this purpose. When you have aligned your assembled sequences, gaps should only be found on the part of the transcripts. If you find genomic sequences which did align to the same transcript and should be overlapping but cannot be aligned to each other you have either found an artifact or an intron which is not completely covered by the genomic sequences. You can have a look onto such sequences by doing a blast so you can estimate if everything looks fine. The first could be done by a script, the latter I would do by hand I guess.