BLAST 2.11 hanging when using local nr database

Question

[scarey]

New to posting: Please be gentle.

I am running a local nr blast with Blastx version 2.11.0. The database was downloaded via update_blastdb.pl.

Trying to annotate my transcriptome, but currently just running a test with a single query.

sudo /opt/ncbi-blast-2.11.0+/bin/blastx -db nr -query test.fasta

If I make my own blast database with other sequences everything works fine, but when I blast against the downloaded nr it just hangs at: Database: All non-redundant GenBank, etc. 331,433,500 sequences; 119,961,264,880 total letters

My test fasta file is formatted fine (and works with other local database). Not sure what I am doing wrong here, any suggestion would be wonderful.

[GenoMax]

You need tens of GB of RAM (~50GB) for blast against nr. How much memory are you allocating to this job?

[scarey]

Our system has 128GB. Do I need to specifically allocate with blastx? I assumed it would use what is available.

Answer 1

[Mensur Dlakic]

Most likely you need to wait longer. It takes a while to load in 330 million sequences, and then it also takes time to search against that database and to create all the alignments.

To speed things up: 1) specify a meaningful E-value as calculating all alignments for E=10 takes time (say, -evalue 0.01); 2) if you don't need all alignments, specify a number that you need (say, -num_alignments 500); 3) if you need to know only in general what matches your protein but don't need absolutely all protein hits, use a database that is trimmed at 90% identity (for example, UniRef90 at ftp://ftp.uniprot.org/pub/databases/uniprot/uniref/uniref90/ ). This will speed up the search at least twice, possibly more.

PS I suggest you specify an output file (-out blastx_output.txt), as searches against the nr database tend to produce voluminous outputs.

PPs It is almost certain that you don't need to run this command as sudo.