Problems with transcriptome assembler Oases

[CarlElit]

Friends,

I am trying to run Oases for transcriptome assembly. The result is far from expected, so I would like to ask whether I am running it in a right way? Thanks.

Here is my running command:

    python scripts/oases_pipeline.py -m 25 -M 29 -o output -d " -strand_specific -shortPaired data/reads.fa" -p " -min_trans_lgth 100 -ins_length 300"

My library is strand-specific and pair-ended with length 67bp. The reads are shuffled as:

    >0(left_mate_forwarded)
    ACTC...
    >1(right_mate_reverse_complemented)
    TATA...

I got some transcripts, but are far from the transcripts annotated, also far from the result of Trinity. The longest contig from Oases is ~2500bp (vs. ~10000bp from cufflinks and ~6000bp from Trinity). The N50 value is also low. It only reports 20 contigs those cover full-length of some transcripts from Cufflinks (totally ~4000), while Trinity reports ~650.

The dataset I am using is a subset of *S. pombe*. Does it matter?

Could somebody help me point out whether something wrong here? Thanks.

[ssing]

It doesn't look like you did anything wrong. I had similar results; while OASES gives very accurately assembled contigs, it gives many fewer contigs of much shorter length.

[CarlElit]

Thanks, @ssing.

Tried to ask the authors. They told me to run with a wider range k-mer values.