Convert Stockholm format MSAs to more efficient a3m format before emitting to publishdir

[tlitfin-unsw]

Description of feature

The stockholm format MSAs are extremely bulky files which leave a large disk footprint - particularly the uniprot.sto files generated for pairing sequences in AlphaFold2 multimer. It might be nice to convert MSA files to a3m format before emitting to publishdir.

[tlitfin-unsw]

Maybe a compressed archive is a better solution to preserve the original data format.

[jscgh]

A quick fix we used is adding a compression step to afterScript. A similar script could be added to the module's main.nf

    withName: 'RUN_ALPHAFOLD2_MSA|RUN_ALPHAFOLD2' {
        afterScript = """
            find . -type f -name '*.sto' -exec zstd -19 --rm {} \\;
        """
    }