Genome in a Bottle NA12878 Human Genome 38 remapped validation set

Scripts to remap the Genome in a Bottle NA12878 validation variant calls to build 38 (GRCh38/hg38) of the human genome.

These convert the VCF calls and assessment region BED files from build 37 to build 38 coordinates using remapping. We take multiple remapping approaches for testing purposes:

• The NCBI remapping service
• CrossMap with UCSC chain files

Usage

Download the inputs with:

cd inputs && bash get_inputs.sh

Run the remapping with:

bash run.sh

Requirements

This depends on external tools to do the actual work:

• Python with pyfaidx
• Perl with XML::XPath
• CrossMap -- can be installed with conda
• GNU parallel
• bedtools
• bcftools
• vt
• GATK MIT licensed scripts
• vcf2bed.py from vcflib

The easiest way to install the Python dependencies is with Miniconda. Then do:

conda install -c bcbio crossmap pyfaidx

Contributors

Deanna Church Brad Chapman