blast+ on Representative_Genomes db

Question

[prostoesh]

Hi everyone! i'm pretty sure it was discussed a million times before, but i can't manage to look up some genes using stanalone blastp.

i've downloaded Representative_Genomes db (its bacterial/archaeal genomes database) so now i have 6 files like Representative_Genomes.00.tar.gz Representative_Genomes.01.tar.gz Representative_Genomes.02.tar.gz Representative_Genomes.03.tar.gz Representative_Genomes.04.tar.gz Representative_Genomes.05.tar.gz

what command should i use to find ~50 top hits for a specific gene? i'm going to do this automatically for a lot of genes and then parse resulting 50 hits

Answer 1

[GenoMax]

Best place to start familiarizing yourself with command line blast are these two NCBI manuals: http://www.ncbi.nlm.nih.gov/books/NBK279690/ and http://www.ncbi.nlm.nih.gov/books/NBK279675/

Unzip and untar the pre-made indexes you have downloaded in a directory. It should results in several files. They all need to stay together.

Run blastn -help to list the command line options for blastn. A rough outline command (you would need to make many selections both for running the program/selecting output etc) you would use would be

$ blastn -db /path_to_dir/Representative_Genomes -query your_gene.fasta -out blast_results_file -max_target_seq 50 -outfmt n (use a real number in place of n to choose one of the tabular output options to be able to parse easily)

If you have protein sequence as your starting point then you would need to adjust the blast strategy (tblastn). "Representative_Genomes" database is for nucleotides.

Answer 2

[GenoMax]

You may be better off getting the protein sequences for all "RefSeq" bacterial genomes from NCBI and making a custom database from those yourself. See this thread for help (@5heikki's answer): A: Download All The Bacterial Genomes From Ncbi You would get all the "faa" files (instead of fna).

Then you could use blastp directly without having to worry about translations etc. References to blast guides are in the answer above.

[prostoesh]

yeah - seems like i downloaded the wrond db :) should have be more alerted when blastp couldn't find Representative_Genomes db in current dirr

thats for your clues - i'm trying all of them now