This directory contains the two (male and female) human genome reference assemblies used by the 1000genomes project. These are in gzipped, fasta format. In addition, two index files are provided for the following purpose. When you convert a SAM file into a BAM file with "samtools import" , you need to provide a ref_list file, which contains the reference name and the length of each reference, one per line for each distinct reference. It is this file that specifies the sort order of the bam file. You can build your own by hand, or generate one as follows: cat human_b36_male.fa | razip -c > human_b36_male.fa.gz samtools faidx human_b36_male.fa.gz (razip comes with the samtools package) The fai file is the index into the compressed reference sequence. The caveat is that if you want to merge two bam files, they need the same sort order, hence we are providing the fai file and the reference that they were built from.