BLAST

Create a DNA blastdb

In order to compare query sequences against reference sequences, you must create a blastdb of your reference(s). This is done using makeblastdb which is included when you install blast.

makeblastdb -in <input fasta> -dbtype nucl -out <label for database>

So if you had a file reference.fasta containing the following records:

>reference_1
ATCGATAAA
>reference_2
ATCGATCCC

You would run the following:

makeblastdb -in reference.fasta -dbtype nucl -out my_database

This would create the following files:

my_database.nhr
my_database.nin
my_database.nsq

Note, the database files are labelled with the -out argument.

Extract fasta sequences from a nucl blastdb

You can extract fasta sequence from a blastdb constructed from a fasta file using blastdbcmd which should be installed when you install makeblastdb.

blastdbcmd -entry all -db <database label> -out <outfile>

If you had a database called my_database which contained the files my_database.nhr, my_database.nsq, my_database.nin and you wanted your fasta output file to be called reference.fasta you would run the following:

blastdbcmd -entry all -db my_database -out reference.fasta

Install blast on ubuntu

apt-get install ncbi-blast+

You can check the version that will be installed in advance here:

https://packages.ubuntu.com/xenial/ncbi-blast+

Extract GI and taxid from blastdb

Data can be extracted from a blastdb using blastdbcmd which should be included in a blast installation. You can specify from the options below as part of -outfmt what metadata to include and in what order.

From the man page:

 -outfmt <String>
   Output format, where the available format specifiers are:
       %f means sequence in FASTA format
       %s means sequence data (without defline)
       %a means accession
       %g means gi
       %o means ordinal id (OID)
       %i means sequence id
       %t means sequence title
       %l means sequence length
       %h means sequence hash value
       %T means taxid
       %X means leaf-node taxids
       %e means membership integer
       %L means common taxonomic name
       %C means common taxonomic names for leaf-node taxids
       %S means scientific name
       %N means scientific names for leaf-node taxids
       %B means BLAST name
       %K means taxonomic super kingdom
       %P means PIG

The example snippet shows how gi and taxid can be extracted from blastdb. The NCBI 16SMicrobial (ftp) blastdb was chosen for this example:

# Example:
# blastdbcmd -db <db label> -entry all -outfmt "%g %T" -out <outfile>
blastdbcmd -db 16SMicrobial -entry all -outfmt "%g %T" -out 16SMicrobial.gi_taxid.tsv

Which will produce a file 16SMicrobial.gi_taxid.tsv that looks like this:

939733319 526714
636559958 429001
645319546 629680