What's BigBWA about?
BigBWA is a tool to run the Burrows-Wheeler Aligner--BWA on a Hadoop cluster. The current version of BigBWA (2.1, november 2016) supports the following BWA algorithms:
- BWA-MEM
- BWA-backtrack
- BWA-SW
All of them work with paired and single-end reads.
If you use BigBWA, please cite this article:
José M. Abuin, Juan C. Pichel, Tomás F. Pena and Jorge Amigo. "BigBWA: approaching the Burrows–Wheeler aligner to Big Data technologies". Bioinformatics 31(24), pp. 4003-4005, 2015.
A version for Apache Spark is available here.
Structure
Since version 2.0 the project keeps a standard Maven structure. The source code is in the src/main folder. Inside it, we can find two subfolders:
- java - Here is where the Java code is stored.
- native - Here the BWA native code (C) and the glue logic for JNI is stored.
Getting started
Requirements
Requirements to build BigBWA are the same than the ones to build BWA, with the only exception that the JAVA_HOME environment variable should be defined. If not, you can define it in the /src/main/native/Makefile.common file.
It is also needed to include the flag -fPIC in the Makefile of the considered BWA version. To do this, the user just need to add this option to the end of the CFLAGS variable in the BWA Makefile. Considering bwa-0.7.15, the original Makefile contains:
CFLAGS= -g -Wall -Wno-unused-function -O2
and after the change it should be:
CFLAGS= -g -Wall -Wno-unused-function -O2 -fPIC
Additionaly, and as BigBWA is built with Maven since version 0.2, also have it in the user computer is needed.
Building
The default way to build BigBWA is:
git clone https://github.com/citiususc/BigBWA.git
cd BigBWA
mvn package
This will create the target folder, which will contain the jar file needed to run BigBWA:
- BigBWA-2.1.jar - jar file to launch with Hadoop.
Configuring
Since version 2.0 there is no need of configuring any Hadoop parameter. The only requirement is that the YARN containers need to have at least 7500MB of memory available (for the human genome case).
Running BigBWA
BigBWA requires a working Hadoop cluster. Users should take into account that at least 7500MB of free memory per map are required (each map loads into memory the bwa index). Note that BigBWA uses disk space in the Hadoop tmp directory.
Here it is an example of how to run BigBWA using the BWA-MEM paired algorithm. This example assumes that our index is stored in all the cluster nodes at /Data/HumanBase/ . The index can be obtained with BWA, using "bwa index".
First, we get the input Fastq reads from the 1000 Genomes Project ftp:
wget ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/data/NA12750/sequence_read/ERR000589_1.filt.fastq.gz
wget ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/data/NA12750/sequence_read/ERR000589_2.filt.fastq.gz
Next, the downloaded files should be uncompressed:
gzip -d ERR000589_1.filt.fastq.gz
gzip -d ERR000589_2.filt.fastq.gz
and prepared to be used by BigBWA:
python src/utils/Fq2FqBigDataPaired.py ERR000589_1.filt.fastq ERR000589_2.filt.fastq ERR000589.fqBD
hdfs dfs -copyFromLocal ERR000589.fqBDP ERR000589.fqBDP
Finally, we can execute BigBWA on the Hadoop cluster:
yarn jar BigBWA-2.1.jar com.github.bigbwa.BigBWA -D mapreduce.input.fileinputformat.split.minsize=123641127
-D mapreduce.input.fileinputformat.split.maxsize=123641127
-D mapreduce.map.memory.mb=7500
-w "-R @RG\tID:foo\tLB:bar\tPL:illumina\tPU:illumina\tSM:ERR000589 -t 2"
-m -p --index /Data/HumanBase/hg19 -r ERR000589.fqBDP ExitERR000589
Options:
- -m - Sequence alignment algorithm.
- -p - Use paired-end reads.
- -w "args" - Can be used to pass arguments directly to BWA (ex. "-t 4" to specify the amount of threads to use per instance of BWA).
- --index index_prefix - Index prefix is specified. The index must be available in all the cluster nodes at the same location.
- The last two arguments are the input and output HDFS files.
If you want to check all the available options, execute the command:
yarn jar BigBWA-2.1.jar com.github.bigbwa.BigBWA -h
The commands are:
BigBWA performs genomic alignment using bwa in a Hadoop/YARN cluster
usage: yarn jar --class com.github.bigbwa.BigBWA BigBWA-2.1.jar
[-a | -b | -m] [-h] [-i <Index prefix>] [-n <Number of
partitions>] [-p | -s] [-r] [-w <"BWA arguments">]
<FASTQ file> <SAM file output>
Help options:
-h, --help Shows this help
Input FASTQ reads options:
-p, --paired Paired reads will be used as input FASTQ reads
-s, --single Single reads will be used as input FASTQ reads
BWA algorithm options:
-a, --aln The ALN algorithm will be used
-b, --bwasw The bwasw algorithm will be used
-m, --mem The MEM algorithm will be used
Index options:
-i, --index <Index prefix> Prefix for the index created by bwa to use - setIndexPath(string)
Spark options:
-n, --partitions <Number of partitions> Number of partitions to divide input - setPartitionNumber(int)
Reducer options:
-r, --reducer The program is going to merge all the final results in a reducer phase
BWA arguments options:
-w, --bwa <"BWA arguments"> Arguments passed directly to BWA
After the execution, to move the output to the local filesystem use:
hdfs dfs -copyToLocal ExitERR000589/part-r-00000 ./
In case there is no reducer, the output will be split into several pieces. In order to put it together users could use one of our Python utils or "samtools merge":
hdfs dfs -copyToLocal ExitERR000589/Output* ./
python src/utils/FullSam.py ./ ./OutputFile.sam
Frequently asked questions (FAQs)
1. I can not build the tool because jni_md.h or jni.h is missing.
You need to set correctly your JAVA_HOME environment variable or you can set it in Makefile.common.
Info
-
- Researchers
- Juan Carlos Pichel Campos
- Tomás Fernández Pena
- José Manuel Abuín Mosquera
- Jorge Amigo Lechuga