BBtools

A suite of command line utilities for sequence quality control, mapping, assembly, and kmer based analysis

Written by Brian Bushnell.

As the BBtools website describes it:

• bbduk – filters or trims reads for adapters and contaminants using k-mers 
• bbmap – short-read aligner for DNA and RNA-seq data 
• bbmerge – merges overlapping or nonoverlapping pairs into a single reads 
• reformat – converts sequence files between different formats such as fastq and fasta
	

For more information see the BBtools website and the BBtools user guide. Support is available from a number of threads at the SeqAnswers site. A listing of those threads is provided here.

The package contains 125 scripts and programs but this examples folder highlights some of the most common use cases.

Directory contents

Running notes

Detailed tutorial instructions

Since we are going to be running computations lets request an interactive node from the job scheduling system. ```bash srun –pty -p short -t 01:00:00 -n 16 -N 1 /bin/bash -I

From there you can examine the output of the results directories for each example. These results are also available at the original location for reference.

First, start by “sourcing” the env.sh script. Doing this this sets up a variable called $bbtoolsExamplesDir that tells the example scripts where the bbtools examples directory is located.

Since we are going to be running these example scripts one-by-one lets remove all the data that is currently in the results directories. Running a helper script can accomplish this easily for you

Trimming adapters

A typical workflow for processing Illumina sequencing data takes the demultiplexed data from the sequencer and perform a series of trimming and contaminant removal steps. The first step is usually to remove adaptor sequences. These adaptors sometimes occur in sequences because the insert size is less than read length and the sequencer reads through to the other side.

This can be accomplished by running this script.

Lets look at the contents of that script.

First it checks for the environment variable $bbtoolsExamplesDir

Then it loads bbtools from the module management system of CERES

Finally it runs bbduk.sh, a program in the bbtools suite:

bbduk.sh can read and write compressed or uncompressed formats and infers the format from the file extension youprovide it.

This will write out a fastq file with adaptors trimmed off.

Kmer filtering for contaminant sequence removal

The next step in quality control is often to remove reads with other contaminants. Typically the file generated from adaptor trimming will be run through bbduk in contaminant filtering mode to remove common lab contaminants like phiX174. For simplicity, this example will use the original input file though, not the output of adaptor trimming.

Run the kmer filtering script.

This calls bbduk.sh with these parameters.

The options are:

Quality trimming

BBduk can perform hard quality trimming. Quality trimming was a pretty standard preprocessing step a few years ago but now the decsion to use it depends on the downstream workflow. Some merging, error correction mapping and assembly methods can use, correct or ignore low quality bases.

Run the adaptor trimming script.

This calls bbduk.sh with these parameters.

The parameters are pretty simple for this process.

Histogram generation

BBduk will generate histogram data for a number of quality metrics. Most of these reports can be generated during another step like adaptor trimming. This example generates many of these reports in one run.

Run the histogram generation script.

This calls bbduk.sh with these parameters.

Each option generates the following histogram data files:

Merging reads

For some applications like assembly, merging reads can reduce the memory required and reduce assembly errors. Running merging also produces an estimate of the read-pair insert size which can help the lab optimize their fragmentation and library creation methods.

Run the basic read merging script.

This calls bbmerge.sh with these parameters.

The parameters are:

Merging for error correction

Overlapping reads can be used to correct sequencing errors on the ends of the read pairs.

Run the histogram generation script.

This calls bbmerge.sh with these parameters.

The parameters are:

Mapping Reads to a reference with bbmap

The reads in this tutorial come from a mock viral metagenome with isolate virus DNA mixed together. The reference database Ref_database.fq.gz used for mapping has those reference viral genomes placed into one file.

Run the basic read merging script. This will build an index and map reads to it.

This calls bbmerge.sh with these parameters. First it builds an index:

Next, it maps the reads to the reference.

The parameters are: