BLASTing masssive protein sequences

dena.dinesh · 01-30-2015, 05:42 AM

Hi,

I have a transcriptome which consists of 38,000 protein sequences. Is it possible to BLAST all of them and find othologs and of these proteins and GeneOnology and pathways involved if there is any for these proteins.

Kindly guide me

Regards
Dena

GenoMax · 01-31-2015, 01:09 PM

It is certainly possible to do the BLASTing.

Question is do you have access to a cluster/adequate hardware and patience, since this will likely take a while to complete. Based on the number of proteins I assume this is a eukaryotic (non-model?) organism. If there is a closely related genome/proteome available then you could significantly narrow your search space and speed up BLAST part.

dena.dinesh · 01-31-2015, 01:21 PM

Hi GenoMax,

Thank you very much for your reply. Yes the organism in which I am working is a eukaryotic non-model organism. I know an organism called "Schistosoma mansoni" is a closely related organism which has genome in NCBI.

I have desktop system which has 256 GB RAM and 16 cores. How long will it take to BLAST the 68000 nucleotides sequence against database and filter out top3-5 homologs and their corresponding Gene Ontology or pathways?
Kindly guide me

Zapages · 01-31-2015, 03:03 PM

I would check this answer and try the Diamond BLASTx option as that will be the quickest route with similar results as NCBI BLAST/BLAST+.

Link to something similar: http://ask.iplantcollaborative.org/q...gg-annotation/

All the best.

westerman · 02-02-2015, 07:06 AM

While Diamond is much faster than BlastX the results are different in that correspondence between the top hit from each program is likely to not be the same. I *think* that this is due to differences in determining the evalues. Digging deeper and taking, say, the top 10 hits will start showing similarities between the programs. This is not to say that Diamond's results are biologically incorrect but you do have to be aware that the programs are not drop-in replacements for each other.

FastAnnot · 02-17-2015, 08:04 PM

It's a pretty big job you're talking about because blastx is exceedingly slow and the data for annotating GOs is also rather cumbersome. For this reason we launched a cloud service fastannot.com which does this task quickly and cheaply. As the previous reply mentioned regarding Diamond, our aligner is also not identical to blastx, but the annotation results are comparable.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
COG categories of protein sequences	Antony03	Bioinformatics	4	12-16-2014 03:46 AM
From a list of protein accessions, how can I retrieve DNA sequences from GenBank?	Ganesh Ujwal	Bioinformatics	0	12-07-2014 07:33 AM
160 protein sequences	tamarbsor	Bioinformatics	1	11-25-2012 11:50 AM

01-30-2015, 05:42 AM	#1
dena.dinesh Member Location: Europe Join Date: Feb 2013 Posts: 58	BLASTing masssive protein sequences Hi, I have a transcriptome which consists of 38,000 protein sequences. Is it possible to BLAST all of them and find othologs and of these proteins and GeneOnology and pathways involved if there is any for these proteins. Kindly guide me Regards Dena

01-31-2015, 01:09 PM	#2
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,143	It is certainly possible to do the BLASTing. Question is do you have access to a cluster/adequate hardware and patience, since this will likely take a while to complete. Based on the number of proteins I assume this is a eukaryotic (non-model?) organism. If there is a closely related genome/proteome available then you could significantly narrow your search space and speed up BLAST part.

01-31-2015, 01:21 PM	#3
dena.dinesh Member Location: Europe Join Date: Feb 2013 Posts: 58	Hi GenoMax, Thank you very much for your reply. Yes the organism in which I am working is a eukaryotic non-model organism. I know an organism called "Schistosoma mansoni" is a closely related organism which has genome in NCBI. I have desktop system which has 256 GB RAM and 16 cores. How long will it take to BLAST the 68000 nucleotides sequence against database and filter out top3-5 homologs and their corresponding Gene Ontology or pathways? Kindly guide me

01-31-2015, 03:03 PM	#4
Zapages Member Location: NJ Join Date: Oct 2012 Posts: 97	I would check this answer and try the Diamond BLASTx option as that will be the quickest route with similar results as NCBI BLAST/BLAST+. Link to something similar: http://ask.iplantcollaborative.org/q...gg-annotation/ All the best.

02-02-2015, 07:06 AM	#5
westerman Rick Westerman Location: Purdue University, Indiana, USA Join Date: Jun 2008 Posts: 1,104	While Diamond is much faster than BlastX the results are different in that correspondence between the top hit from each program is likely to not be the same. I think that this is due to differences in determining the evalues. Digging deeper and taking, say, the top 10 hits will start showing similarities between the programs. This is not to say that Diamond's results are biologically incorrect but you do have to be aware that the programs are not drop-in replacements for each other.

02-17-2015, 08:04 PM	#6
FastAnnot Junior Member Location: USA Join Date: Dec 2014 Posts: 4	It's a pretty big job you're talking about because blastx is exceedingly slow and the data for annotating GOs is also rather cumbersome. For this reason we launched a cloud service fastannot.com which does this task quickly and cheaply. As the previous reply mentioned regarding Diamond, our aligner is also not identical to blastx, but the annotation results are comparable.