assemble RNAseq data with Oases but so many transcripts in locus_1

[minoru_harvest]

Hello everyone

I'm doing transcriptome assembly work with Oases.
The illumina data's read length is 100nt and its quality is pretty good. I filterred bed reads, trimmed bases of low quality and trimmed adapters. Then did the Oases_pipeline in this command:

nohup oases_pipeline.py -m 17 -M 71 -s 2 -g 27 -o Hg0912 -d " -short Hg-trim.fasta " &

The final result in transcripts.fa seems not so good. The first locus "locus_1" has so many transcripts, which exceeds 600,000.

>Locus_1_Transcript_1/656445_Confidence_0.000_Length_257
TTATTTTCTTCCTGTTGTTTTCAGTACGAGCCAGTTGAGATGCGCGTGAGTTTATAAACA
AAACCTGTGTCCCCGATTGGCCAGTAAGTAGCCGGCAACCGACACGGACGTTGTACTTGT
ATTGAGCAAAGTTTATTCCAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACA
TCCCGACGAATTGAGAACCTCTCCTTTTTGGGATAAAAAAAAAAAAAAAAAAAAAGTTGT
ATTTTGTGTTTCAAAGT
>Locus_1_Transcript_2/656445_Confidence_0.000_Length_505
CATTCCTTGTATTCAAAAAAAAAAAAAAAAAAAAAAAAATGAAACACGTCAACAAAAAAA
AAAAAAGGAACCCTTATTCTTAGAGAATTAAGACTTTTGCAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAACATCCCGACGAATTGAGAACCTCTCCTTTTTGGGATAAAAAAAAAAA
AAAAAAAAAAGCATGAAAAAAAAAAAATGCAGAAATCCCACTTTACTTTTAGATAAATAT
TGCAAATTTGCGATACATAACATACATTAATTACATATAGGTAACTGTTTATTTTAAGGC
AAATTCTTAGAAAAAACTAAGAAGTCCTGGATCAACTAAAAAATACAGCTCTCGAACGTC
GCTCTTACAATTTTAAAACCAAGTTCCTTGAGTTAAAAATTGGAAAAAGTCGCGCTCGCT
CCGCTCGCGATTTAGAAGCGATGTGCTTGTTTTTGCATTCGCCGGCCAACCAACAAAAAA
TTATGGACGTTTGAGCTACACTTAT
>Locus_1_Transcript_3/656445_Confidence_0.000_Length_407
TCAAAAAAAAAAAAAAAAAAAAAAAAATGAAACACGTCAACAAAAAAAAAAAAAGGAACC
CTTATTCTTAGAGAATTAAGACTTTTGCAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAC
ATCCCGACGAATTGAGAACCTCTCCTTTTTGGGATAAAAAAAAAAAAAAAAAAAAAGCAC
GTCAACAAAAAAAAAAAAAGGAACCACATTGAATTTTGTGTTACGTCACTACTTTTAGGC
......

And I pick out a long sequence to do blastx ,nothing similar found. But when blastn, the sequence can match to my species' genome sequences(my species do not have a complete genome sequneced, the matched subjects are something like BAC library).

Is this normal?

[mariruilo]

Hi Minoru,

I'm having the same problem. Did you find an answer for it? Thanks.

[Cofactor Genomics]

Hi,
Although our team has not seen this many transcripts under a given locus... what we have tended to do is set those loci with a given blastx hit as tier one transcripts (ie given them a higher ranking/priority over the other transcripts).

On another note, not necessarily related to a scenario of the 6,000 transcripts under a single locus, but focussed on the results coming out of a transcriptome assembly: We found that the molecular biology protocols and depletions have a large impact on the resulting transcriptome assembly. In the early days before we had the protocols nailed down we found that a very large percent of our loci did not have blastx hits, as our molecular biology (ie not our bioinformatics) became more refined, we saw a much higher percent of loci matching to genes.

I hope this is helpful.

Jarret Glasscock
Cofactor Genomics

http://www.cofactorgenomics.com

[mariruilo]

Thank you very much for the answer, it is very helpful.

Mari