Version

• NCBI BLAST version 2.2.20 is available in murska.csc.fi
  
• NCBI BLAST version 2.2.20 is available in hippu.csc.fi.
  
• An optimized blatstpgp (PSI-BLAST) is available in murska.csc.fi.
• A WWW interface to BLAST is available in the Scientist's Interface.

Description

BLAST (Basic Local Alignment Search Tool) is the most frequently used sequence homology search tool. Given a probe sequence (nucleotide or protein), BLAST compares it to a sequence database and picks out sequences with significant similarity to the probe sequence. BLAST uses a heuristic search protocol, which makes search very fast compared to non-heuristic methods. The heuristics used may however cause BLAST to fail to find all significant hits.

The command line version of NCBI-BLAST allows a user to modify all parameters of BLAST, to use special methods like PSI-BLAST and PHI-BLAST, and to analyze large data sets.

In Murska and Corona you can also use pb (Parallel Blast) command for large sets of query sequences. The pb program splits a large search jobs into severall subjobs, that can be executed simultaneously (more below).

NCBI BLAST includes e.g. the following programs

• blastall: normal BLAST searches (blastp, blastn, blastx, tblastn, or tblastx)
• blastclust: clusters protein or DNA sequences based on pairwise matches found using the BLAST or MegaBLAST algorithm.
• blastpgp: performs gapped blastp searches and can be used to perform iterative PSI-BLAST and PHI-BLAST searches.
• bl2seq: performs a comparison between two sequences using the blast algorithm.
• megablast: uses the greedy algorithm of Webb Miller et al. for nucleotide sequence alignment search and concatenates many queries to save time spent on scanning the database.

Usage of NCBI BLAST

WWW interface

Two BLAST WWW interfaces are available in the Scientist's Interface. One for searcing from the public databases and one for
searching form users own sequence set.

Command line usage in Hippu.csc.fi

To set up NCBI BLAST in hippu.csc.fi, give command:

module load blast

After this you can check the command line options or a specific BLAST program by typing program_name --help, for example:

blastall --help

Example of command line BLAST usage:

BLAST search with fasta-formatted protein sequence file "protsequence.fasta" against PDB database:

blastall -p blastp -d pdb -i protseq.fasta -o results.out

Hippu.csc.fi is a very effective platform for doing BLAST searches  and it has plenty of memory and several multicore processors that BLAST can utilize. To speed up your search you can select to use more than 1 processor for compting. This is done with option -a. For example that dommand bellow does the search using 8 processors.

blastall -p blastp -a 8 -d pdb -i protseq.fasta -o results.out

Usage of pb (Parallel BLAST) in Murska and Hippu

If your query sequce set contains less than 20 sequcenses the hippu is probable the most effective platform for the search. However, if your query set contains hundreds or thousands of sequenses then utiizing the murska.csc.fi cluster environent is more  effective. For this kind of massive blast searches you can also utilize the pb commad.

In Murska and Hippu , you can execute any blastall, megablast or blastpgp command through the pb command. pb (Parallel BLAST) is designed for situations, where the query file includes several sequences. In murska.csc.fi It splits the query task into severall subjobs, that can be run simultaneously using the resources of the server very effectively. For large sets of query sequences, pb can speed up the search up to 50 fold. Two sample pb commands for murska.csc.fi:

module load blast

pb blastall -p blastn -d embl -i 100_ests.fasta -o results.out

pb blastpgp -d swiss -i protseq.fasta  -j 3 -a 4 -o results.out

The pb program also allows users to do BLAST searches against their own fasta formatted sequence sets. This is done by replacing the -d option with option -dbnuc (for nucleotides) or -dbprot (for proteins). This utility is useful in hippu.csc.fi too.  Example: 

pb blastall -p blastn -dbnuc my_seq_set.fasta -i querys.fasta -o results.out

Whit pb, you can also use three addintional output formats for BLAST results.
-m 12 prints out the hit sequences as EMBOSS list file
-m 13 prints out the hit sequences in fasta format
-m 14 prints out the matching regions of the hit sequences in fasta format

BLAST databases in Murska and Hippu

Name	database	source files
Nucleotides
arabidopsisN	Aabidopsis thaliana nucleotide sequences	ArabidopsisN.Z
embl_est	EMBL EST division	est files
embl_gss	EMBL GSS division	gss files
embl_htg	EMBL HTG division	htg files
embl_others	EMBL excluding the EST, GSS and HTG divisions, and the WGS sets	other files
emblnew	EMBL updates	cum files
nt	NCBI non-redundant nucleotide database	nt.gz
refseq	NCBI RefSeq RNA database	rna.fna.gz files
refseq_con	NCBI RefSeq human contigs	all chromosomes
Proteins
nr	NCBI non-redundant protein database	nr.gz
pdb	PDB protein structure database	pdb_seqres.txt
swiss	Uniprot/Swiss database	uniprot_sprot.fasta.gz
trembl	Uniprot/TrEMBL database	uniprot_trembl.fasta.gz
uniref100	Uniref100 database	uniref100.fasta.gz
uniref90	UniRef90 database	uniref90.fasta.gz
uniref50	UniRef50 database	uniref50.fasta.gz

Users own databases

Users can do BLAST searches against their own fasta formatted sequence sets with My BLAST
www interface:

• My BLAST
  

Other option is to use pb blast in Murska or in Hippu. In pb blast own databases are used by replacing the -d option with option -dbnuc (for nucleotides) or -dbprot (for proteins). Example: 

pb blastall -p blastn -dbnuc my_seq_set.fasta -i querys.fasta -o results.out

More information

• BLAST manual
• NCBI BLAST server