BLAST sequence searching on Big Red

On this page:

• Publicly installed databases
• Preparing to run BLAST
• Running NCBI blastall interactively
• Running NCBI blastall batch jobs
• NCBI blastall options
• Dealing with large volumes of output
• Manual pages

Publicly installed databases

You can reference databases using the names in the "Database" column (e.g., -d nt).

Preparing to run BLAST

Before you can run BLAST, you need to add it to your path. To permanently add BLAST to your path, add the +ncbi key to your .soft environment. To do so, run:

This places blastall in your path, and sets the environment variables BLASTMAT and BLASTDB.

Running NCBI blastall interactively

To run NCBI blastall interactively from the command line, you can use only one processor and your blastall job must consume less than 20 minutes of processor time. If you need more than 20 minutes of processor time, use the serialjob command (see below) to run blastall.

Running NCBI blastall batch jobs

If your blastall job will run for more than 20 minutes, you must use the batch facility.

Use the serialjob command as a one-step method for submitting blastall jobs to the queue. The serialjob command is a prefix to a standard blastall command. Use arguments just as you would with blastall, with a few optional additions. If your job will take less than two hours, you can run:

Replace options_to_blastall with your command options.

If your job will run for more than two hours, specify the -wallclocklimit option:

Replace hh with hours, mm with minutes, and ss with seconds.

Use the double dash (--) to separate the set of options to blastall from those that will exceed two hours. For example, if you want to blast nucleotide sequences in file sequences.mine against the patent database, and have the output stored in file hits.patent, and have the job run for four hours, run: serialjob blastall -p blastn -i sequences.mine -d patnt -o hits.patent -- -wallclocklimit 4:00:00

NCBI blastall options

Options for blastall are available by entering the command with no argument. Options are listed here for your convenience:

Dealing with large volumes of output

Blast output is verbose. Default output from thousands of searches will consume hundreds of megabytes or gigabytes. To handle large output, you can work using scratch disk volumes and use blastall options that reduce the volume of output:

Using scratch space

Quotas on disk space exist for home directories. UITS will not increase quotas for individual users to many hundreds of megabytes or to gigabytes. To use large amounts of storage space for short periods of time, use scratch space: /N/gpfsbr

To set up a personal scratch directory, use mkdir with your Big Red username:

Scratch space is for temporary use and depends on the honor system. Official policy is any file more than three months old can be deleted without notice. Make sure to backup your data to a long-term storage system.

Minimizing blastall output

Tab-delimited output: If you are querying hundreds or thousands of sequences against databases, you probably will use programs to summarize the output. Blastall offers two options for reducing output to tab-delimited tables for fields that programs can easily read. The blastall option -m 9 produces a table with headers and comment information at the top, and -m 8 produces a table without headers and comments. Fields are output in the following order:

• Identity of query sequence
• Identity of subject sequence (matching sequence in database)
• Percent identity
• Alignment length
• Number of mismatches
• Number of gaps
• Start of query sequence
• End of query sequence
• Start of subject sequence
• End of subject sequence
• E-value
• Bit-score

Manual pages

Unix manual pages provide reference information about the scripts serialjob and parblastjob. To access them, use the man command.

This document was developed with support from the National Science Foundation (NSF) under Grant No. 0503697 to the University of Chicago and subcontracted to Indiana University. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the NSF.

Comments/Questions/Corrections