You are here
DIAMOND - sequence aligner
Table of Contents
Introduction
DIAMOND is a program to do local alignments of nucleotide (BLASTX compatible) and amino acid (BLASTP compatible) sequences to a reference protein sequence database. It has been designed to be much faster than the corresponding BLAST programs. DIAMOND can use multiple cores on a single compute node (using thread-based parallelization). Do not request multiple nodes when running DIAMOND. Instead, use a PBS directive in your job script (or corresponding qsub command line argument) of the form:
#PBS -l nodes=1:ppn=n
where n is the number of cores you would like to use.
Running DIAMOND on Breezy
DIAMOND 0.7.9 was compiled with GNU GCC compilers on Breezy and installed to /global/software/diamond/diamond079/bin. The executable program is called diamond. Since the software was built with a more recent version of the GCC compilers than the default on Breezy, use the module command shown below to initialize the environment for that compiler. For a list of diamond command line arguments use the -h argument:
module load gcc/gcc-4.9-20140406-beta
/global/software/diamond/diamond079/bin/diamond -h
To match the number of threads used to the number requested of the batch job system, use the TORQUE variable PBS_NUM_PPN with the -p flag when calling diamond:
module load gcc/gcc-4.9-20140406-beta
/global/software/diamond/diamond079/bin/diamond -p $PBS_NUM_PPN other_command_line_arguments
For More Information
Page created: 2015-04-09.