v3-alpha.1 - publish genotype cache after PGSC_CALC_LOAD finishes not as part of PGSC_CALC_SCORE

[sritchie73]

Description of feature

If PGSC_CALC_SCORE crashes (e.g. due to not being able to meet resource requirements listed in the config file, see #470 ) then the genotype cache is not published to the output folder for use in future runs despite PGSC_CALC_LOAD completing successfully.

I would suggest moving the cache publishing code out of PGSC_CALC_SCORE to occur after aggregation step in the main pipeline (line 126 of workflows/pgsc_calc/main.nf) so that the genotype cache is published in a format that can be used by --genotype_cache_zip in subsequent runs.

[sritchie73]

For anyone else with this issue, I managed to reverse engineer the process required to combine the individual chromosome genotypes.zarr.zip into a single genotype.zarr.zip file that can be used with --genotype_cache_zip:

#!/bin/bash

# Script must be run from the directory you originally launched the pipeline from

# Find all the genotype.zarr.zip files for the individual chromosomes
chr_caches=$(find work -name 'genotypes.zarr.zip')

# Unzip them, in parallel, creating a new combined genotypes.zarr folder in
# the current directory
for ff in $chr_caches; do 
  (
    7z x -y $ff
  ) &
done

# Zip the result up to create the combined archive
7z a -tzip -mx0 genotypes.zarr.zip genotypes.zarr

# Remove folder 
rm -rf genotypes.zarr

# Move combined archive to desired output location
mv genotypes.zarr.zip ${HOME}/out/