Blastp or blastx for functional annotation after Trinity assembly?!

[setar...@gmail.com]

Hi all,

I have done a transcriptome assembly using Trinity. To facilitate the functional annotation process and save much time, I used Transdecoder to ORF prediction of assembled transcripts, and so I used blastp for blasting the deduced amino sequences against multiple databases. Although blastx actually is ORF finding in 6 frame + blastp, most papers use blastx not blastp that a bit worried me for reviewing and publication. As far as I know Transdecoder finds ORF in all 6 frames and returns all of them, not the best one yes? so using this tool along with blastp doesn't sound problematic, am I right? Please relief me with your helpful comments and experiences.

Thank you in advance,

[Brian Haas]

Hi,

If you're using the final Transdecoder.pep file, then it's not all orfs, but just those that 'look' like coding regions based on some coding statistics.

I'd suggest following our Trinotate protocol, which includes the blastp and blastx steps:

http://trinotate.github.io/

This way, you'll capture homology for those regions that have short coding regions are otherwise didn't meet the coding metrics criteria in TransDecoder.

best,

~b

--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-u...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.

--

--
Brian J. Haas
The Broad Institute
http://broadinstitute.org/~bhaas

 

[setar...@gmail.com]

Hi Brian,

Many thanks for your prompt response. Yes, I used the Transdecoder.pep file, and satisfied with blast results. Actually, about 80-85% of protein sequences were got best hit from NR, UniProt. This result shows that the potential protein sequences were correctly predicted by Transdecoder, am I right? Since I see most papers use blastx (not blastp), I want just to know if  this way (Transdecoder.pep file for blastp) is reasonable and if the blastx has any superiority in relative to blastp?

Thank you

[Brian Haas]

Blastp is very fast and will give you what you want when TransDecoder found it, but TransDecoder isn't perfect, and you'll still get useful info from running a blastx too.  So, TransDecoder/BlastP  is not an outright replacement for running BlastX.   But, you have freedom to do whatever you want. ;)

best,

~b

--
You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-u...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.

[setar...@gmail.com]

Thank you, Brian. Yes, neither Transdecoder nor any ORF prediction tool is perfect. However, I did blastp instead of blastx just for increasing the speed; I could get the reasonable results in a short time. From the point of a reviewer, please let me know if you accept my procedure?

Kind regards,

On Tuesday, May 17, 2016 at 7:42:29 PM UTC+4:30, setar...@gmail.com wrote:

4

[Brian Haas]

It all depends on what sort of claims you're going to make. 

If you find an interesting transcript and want to report on it, and you say it doesn't have a presumed function since there was either no ORF, or no blastp hits to some predicted ORF, I would ask - why don't you do a blastx and see if you find any potential homologs. 

The rigor of the analysis should reflect what you're aiming to accomplish.  (my $0.02)

~b

--

You received this message because you are subscribed to the Google Groups "trinityrnaseq-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trinityrnaseq-u...@googlegroups.com.
To post to this group, send email to trinityrn...@googlegroups.com.
Visit this group at https://groups.google.com/group/trinityrnaseq-users.
For more options, visit https://groups.google.com/d/optout.