Re: [Transdecoder-users] REgd: Transdecoder in Transcriptome assembly

[Brian H. <bh...@br...>]

Hi,

For transdecoder to work, it needs the full sets of transcripts to operate
on, from which it then builds a statistical model and ultimately goes
through the full set of transcripts and predicted ORFs to find those that
look to be coding based on the model.  Running it on just a couple of
transcripts isn't going to produce useful results, is my guess. Also, the
ORF must be sufficiently long (min length is 100 aa or 300 nts) and not
frameshifted or otherwise interrupted.

Please use the transdecoder user support mailing list (CC'd) for user
support.

best wishes,

~brian

On Tue, Jan 13, 2015 at 2:47 AM, Srilakshmy Lakshminarayanapuram <
sri...@sl...> wrote:

>  Hello Brian,
>
>
>  I am a Master student working on my thesis right now with RNA seq data.
> I am stuck with the use of Transdecoder at this point of time.
>
> I did an assembly of the transcriptome with Trinity and transabyss. My
> goal is to identify paralogs in the assembly of a reported whole genome
> duplication event.
>
> I work on a non-model species.
>
> My workflow so far is:
>
> 1. Assembled transcriptome (~200,000 contigs)
>
>
>  2. Filtered based on expression and length (FPKM :1 and length of the
> transcript > 500)
>
>
>  3. Self reciprocal blast search to identify paralogs. Got the pairs of
> best hits from the list.
>
>
>  4.Just to test, applied a very simple nucleotide substitution model and
> calculated K (substitution rate) for the pairs of paralogs identified from
> the blast.
>
>
>  5. We got a very high single peak form K rate and just a little
> conformation that at least paralogs might exists in the assembly.
>
>
>  But to get a clear picture, we wanted to look at peptide level.
>
>
>  So for each pair of paralogs we identified from blastn, i'm trying
> to convert the nucleotide sequence in to peptide. But for   most of the
> pairs, they seem not to convert but just gives an empty file.
>
>
>
> I even tried to convert the whole filtered assembly to peptide, even there
> some of the sequences does not seem to get converted to peptide sequence.
> In few there seem two ORF's predicted.
>
>
>
> My problem is i know they could be paralogs, so to test i need to change
> in to peptide sequence. I will post my query and also an example of
> sequence which is not changed.
>
>
>  Query:
>
> $TRINITY_RNASEQ_ROOT/trinity-plugins/transdecoder/TransDecoder -t $1pep.fa
> --CPU 32
>
> I will attach 1pep.fa for your reference.
>
> Your help will be greatly appreciated.
>
>
>
>  Thanks for your time,
>
> Srilakshmy
>



-- 
--
Brian J. Haas
The Broad Institute
http://broadinstitute.org/~bhaas <http://broad.mit.edu/~bhaas>