Is it possible to run this with 8 GB memory on Mac?

[Fiwx]

Hi, I am on a Mac with 8 GB memory. I was wondering if it is possible to run pgsc_calc, and if so, what parameters to modify to allow this. Thank you.

[nebfield]

You could try 😅 It depends on the size and complexity of your target genomes and scoring files.

First you'd need to increase the amount of memory / CPU in Docker Desktop. The default limits are quite low.

Then just follow the instructions here to change pipeline resources (although instead of using more resources, you want to use less). As a quick example, make a new text file that contains:

process {
    executor = 'local'

    withLabel:process_low {
        cpus   = 1
        memory = 8.GB
        time   = 1.h
    }
    withLabel:process_medium {
        cpus   = 2
        memory = 8.GB
        time   = 4.h
    }
    withLabel:process_high_memory {
        cpus   = 2
        memory = 8.GB
        time   = 4.h
    }
}

and save it as something like custom.config

then run:

$ nextflow run pgscatalog/pgsc_calc -c path/to/custom.config ... (your normal parameters)

When you run the pipeline, if you get any errors with exit code 137 that means you've run out of memory and the workflow was automatically killed by docker 😬 There's no fix for this, so you'd need to find a computer with more memory.

[smlmbrt]

It also depends on how many variants are in your genotyping data (the fewer the better in this case) - plink will error out if it doesn't get >8 Gb for the LD-pruning step.