Transcripts from RNA-seq assembly

[StopCodon]

I assembled the transcriptome of an unsequenced organism using Trinity. I would like to do some downstream analysis on the assembled transcripts. When I look at the transcripts and translate them, I find they are littered with stop codons, does that indicate there is some issue with sequencing? How can one go from the transcript to the amino acid sequence? Can we assume that the beginning of the transcript would be the beginning of the transcription start site for the gene? I would love any pointers on this.

Thanks,

[gringer]

Can we assume that the beginning of the transcript would be the beginning of the transcription start site for the gene?

This is unlikely. Trinity assembles the whole RNA, including untranslated regions. If you want to look at protein predictions de-novo, then you need something like frameDP to correct for frame shift errors and identify the most likely start/stop sites for translation.

[Simon Anders]

Are you sure you were in the right reading frame when checking for stop codons? Look for start codons to fix the reading frame.

Can we assume that the beginning of the transcript would be the beginning of the transcription start site for the gene?

I would say, possibly yes, if you meant to say transcription start site. But gringer might have been right in assuming that you meant the translation start.

[gringer]

I would say, possibly yes, if you meant to say transcription start site. But gringer might have been right in assuming that you meant the translation start.

Whoops, yes, I read that wrong. Just to add to this, depending on the mapping quality the start point may not necessarily be exactly at the transcript start location.

[StopCodon]

Thanks. That clarifies some concepts. Yes, I had meant the Translation start site instead of transcription. Now my data makes more sense as well since very few of the assembled transcripts were starting with a start codon. Is FrameDP a good choice for finding peptide sequences? I saw it had only 10 citations to date. Are there any other tools for doing this?

Thanks.

[gringer]

Is FrameDP a good choice for finding peptide sequences? I saw it had only 10 citations to date. Are there any other tools for doing this?

Have a look here for other options:

http://seqanswers.com/forums/showthread.php?t=4036

FrameDP generates huge amounts of files, something like 5-10 times the number of transcripts, so you need to be careful with that. My computer claimed to have run out of space (with ~300GB free) due to me not being careful enough running that program.

[GillermoPonz]

I'm having the same problem just now. When I translate the assembled transcripts (in the 6 pssible reading frames), they are full of stop codons and I don't know what to do. I've been thinking about not normalizing the libraries when assembling in Trinity but I don't know if it make sense...

Did you finally solve the problem?

Thank you