6.1 Genome indexing for bowtie

To map reads to a reference genome, all mapping tools require the genome to be converted into particular structures for quick access and to keep the memory footprint small. In this tutorial we will use tophat2 for the transcriptome reconstruction, which in turn requires bowtie2 to run for the read mapping to the genome. Bowtie2 needs the genome to be indexed using the BurrowsAWheeler transform, and provides a tool (bowtie2-build) to obtain this transformation starting from the genome sequence stored in a text file in fasta format. You will also try another mapper, STAR, which performs quite well and can be easily integrated with cufflinks/cuffdif, as you will see later.

Let’s first look at the bowtie2-build options:

$ module load bowtie2
$ bowtie2-build --help

The general usage is:

Usage:

bowtie2-build [options]* <reference_in> <bt2_index_base>

You need to specify the genomic sequence file (reference_in) and a label to identify the index (bt2_index_base), which will be the prefix of all files written by bowtie2-build. Copy a folder containing the genomic sequence with the following command:

$ cp –r /pico/home/userexternal/fferre00/tutorial/genome .
 Check the content of the genome folder that you just copied: 

$ ls –l genome