How to liftover scoring files to match your input genome build

How to liftover scoring files to match your input genome build#

pgsc_calc uses genomic coordinates to help match variants in the scoring file to variants in your input genomic data. These coordinates may not match if scoring files are developed in a different genome build. Luckily, it’s fairly simple to automatically remap coordinates (liftover) to different genome builds.

Lifting over PGS Catalog scoring files#

The PGS Catalog provides scoring files in builds GRCh37 and GRCh38. The pipeline queries the PGS Catalog API and automatically downloads the appropriate scoring files to match the input target genome build.

Just set the build of your target genomes with --target_build and pgsc_calc will do the rest.

Lifting over custom scoring files#

Custom scoring files need to include metadata in a header, following the PGS Catalog v2 file format standard. A single line can be included at the top of your custom scoring file to set the genome build:

#genome_build=GRCh37

Valid genome builds are GRCh37 and GRCh38.

Tip

If you’re having trouble getting your scorefile working, see the example we use in our automatic tests

Once your scores have valid headers, the pipeline can automatically liftover scoring files to match your input data if you set the following additional parameters at runtime:

--liftover --target_build GRCh38

Where --target_build can be GRCh37 or GRCh38.

Note

You also need to provide the path to liftover chain files --hg19_chain and --hg38_chain

Putting everything together for an example run, assuming the input genomic data are in build GRCh38:

$ nextflow run pgscatalog/pgsc_calc \
    -profile <docker/singularity/conda> \
    --input samplesheet.csv \
    --scorefile MyCustomFile.txt \
    --liftover \
    --target_build GRCh38 \
    --hg19_chain <path/to/hg19ToHg38.over.chain.gz> \
    --hg38_chain <path/to/hg38ToHg19.over.chain.gz>

Note

If you’re calculating multiple scores (see How to apply multiple scores in parallel) and you want to liftover some of the scorefiles, then all scorefiles need genome builds set in the header