A Ruby gem to generate a per contig score using exonerate and a genome

transonerate --assembly <your_assembly> --genome <genome> --annotation <
existing_annotation> --left <fastq_files> --right <fastq_files>

Idea for getting a per contig score using exonerate: We want to show that the transrate score, which is achieved used reads only, is highly correlated with this exonerate score which requires a genome and gtf.

Align the reads to the genome using tophat/cufflinks and create a new gtf file that contains all the previously annotated regions and also any new ones that it thinks exists based on the read evidence.

From exonerate we can set ryo to

--ryo "%qi\t%ti\t%pi\t%qab\t%qae\t%tab\t%tae\t%ql\n"

that will output a sort of blast6-like format. Eg:

query               target  seqid  qs  qe    target_s   target_e     qlen
comp111716_c0_seq1    Chr1  95.24  0   253   33596559   33596305     253

Then we can find the length of alignment by doing (qe-qs/qlen) and multiply it by the %seqid. Then look up this region in the gtf. Sum the length of all exons that are wholly or partially covered by the target_s and target_e exonerate hit. Find the percentage coverage to the transcript. So the score has to take into account the %seqid, the percentage of the contig that aligned to the genome, and the percentage of the aligned contig that aligned to a transcript in the gtf.