BWA is a fast light-weighted tool that aligns short sequences to a sequence database, such as the human reference genome. By default, BWA finds an alignment within edit distance 2 to the query sequence, except for disallowing gaps close to the end of the query. It can also be tuned to find a fraction of longer gaps at the cost of speed and of more false alignments.
BWA excels in its speed. Mapping 2 million high-quality 35bp short reads against the human genome can be done in 20 minutes. Usually the speed is gained at the cost of huge memory, disallowing gaps and/or the hard limits on the maximum read length and the maximum mismatches. BWA does not. It is still relatively light-weighted (2.3GB memory for human alignment), performs gapped alignment, and does not set a hard limit on read length or maximum mismatches.
Given a database file in FASTA format, BWA first builds BWT index with the 'index' command. The alignments in suffix array (SA) coordinates are then generated with the 'aln' command. The resulting file contains ALL the alignments found by BWA. The 'samse/sampe' command converts SA coordinates to chromosomal coordinates. For single-end reads, most of computing time is spent on finding the SA coordinates (the aln command). For paired-end reads, half of computing time may be spent on pairing (the sampe command) given 32bp reads. Using longer reads would reduce the fraction of time spent on pairing because each end in a pair would be mapped to fewer places.
How to Use
There are multiple versions of BWA available. An easy way of selecting the version is to use modules. To see the modules available, typemodule avail bwa
To select a module, typemodule load bwa/[ver]
where [ver] is the version of choice. This will set your $PATH variable.
Pre-build BWA index files are available in/fdb/igenomes/[organism]/[source]/[build]/Sequence/BWAIndex/genome.fa
- [organism] is the specific organism of interest (Gallus_gallus, Rattus_norvegicus, etc.)
- [source] is the source for the sequence (NCBI, Ensembl, UCSC)
- [build] is the specific genome draft of interest (hg19, build37.2, GRCh37)
Some users have noticed that newer version of BWA doesn't work with index files from previous versions in /fdb/bwa/indexes. Please use the index files above under /fdb/igenomes instead.
Sample Sessions On Biowulf
BWA sample files can be copied from:/usr/local/bwa/sample
As a rule of thumb, assume that bwa will require as much or more memory as the size of the .bwt index file. For example, the hg19 bwa index file is 2.9gb:$ ls -lLh /fdb/igenomes/Homo_sapiens/UCSC/hg19/Sequence/BWAIndex/genome.fa.bwt -rwxr-xr-x 1 maoj staff 2.9G Mar 15 2012 /fdb/igenomes/Homo_sapiens/UCSC/hg19/Sequence/BWAIndex/genome.fa.bwt
If the UCSC/hg19 BWA index file were used, the bwa process will need at least 3gb of memory.
BWA is a multithreaded application. That is, the bwa command can distribute its work across multiple CPUs on a single node. The number of threads BWA will use is controlled by the -t option. The total number of threads allocated to multiple bwa processes on the same node should not exceed the total number of CPUs on the node.
Submitting a single BWA batch job
1. Create a script file. The file will contain the lines similar to the one below.
2. Make sure you use an appropriate number of threads (-t) for bwa processes. For example, g72 nodes have 16 CPUs, while g4 nodes have 2 CPUs. in the example below, bwa is directed to use four threads:#!/bin/bash # This file is runBWA # #PBS -N BWA #PBS -m be #PBS -k oe cd /home/user/bwa/run1 module load bwa bwa index -a bwtsw tttF3.csfasta bwa aln -t 4 tttF3.csfasta ttt.fastq > ttt.sai bwa samse tttF3.csfasta ttt.sai ttt.fastq > ttt.sam
3. Submit the script using the 'qsub' command, e.g.qsub -l nodes=1:g8 /home/username/runBWA
Here the g8 node was requested so in the script -t 4 was used since g8 node has 4 CPUs. Run freen on biowulf to check for core number for different nodes.
Submitting a swarm of BWA jobs
Using the 'swarm' utility, one can submit many jobs to the cluster to run concurrently.
Set up a swarm command file (eg /home/username/cmdfile). Here is a sample filecd /home/user/bwa/run1 ; bwa index -a bwtsw tttF3.csfasta ; \ bwa aln -t 4 tttF3.csfasta ttt.fastq > ttt.sai ; \ bwa samse tttF3.csfasta ttt.sai ttt.fastq > ttt.sam cd /home/user/bwa/run2 ; bwa index -a bwtsw tttF3.csfasta ; \ bwa aln -t 4 tttF3.csfasta ttt.fastq > ttt.sai ; \ bwa samse tttF3.csfasta ttt.sai ttt.fastq > ttt.sam ... cd /home/user/bwa/run15 ; bwa index -a bwtsw tttF3.csfasta ; \ bwa aln -t 4 tttF3.csfasta ttt.fastq > ttt.sai ; \ bwa samse tttF3.csfasta ttt.sai ttt.fastq > ttt.sam
The -f option is required for swarm. Because bwa is multithreaded, the -t option is be used to direct swarm to allocate multiple cpus per bwa process. Also because full genome alignments using bwa require substantial memory utilization, the -g option can be used to direct swarm to allocate how many gb of memory per bwa process.
By default, swarm will execute each line on one CPU, using 1gb of memory. In the above case, bwa requires four threads, so the swarm commandline should be:swarm -f cmdfile -t 4 --module bwa
If a larger BWA index file were used, for example hg19, then the amount of memory per bwa process must be increased using the -g option:swarm -f cmdfile -t 4 -g 3 --module bwa
For more information regarding running swarm, see swarm.html
To see a full listing of the options available for bwa, type bwa at the prompt.$ bwa Program: bwa (alignment via Burrows-Wheeler transformation) Version: 0.6.2-r126 Contact: Heng Li <email@example.com> Usage: bwa
[options] Command: index index sequences in the FASTA format aln gapped/ungapped alignment samse generate alignment (single ended) sampe generate alignment (paired ended) bwasw BWA-SW for long queries fastmap identify super-maximal exact matches fa2pac convert FASTA to PAC format pac2bwt generate BWT from PAC pac2bwtgen alternative algorithm for generating BWT bwtupdate update .bwt to the new format bwt2sa generate SA from BWT and Occ pac2cspac convert PAC to color-space PAC stdsw standard SW/NW alignment