Take out the leftovers

I’m one of the people who has to present a paper at my lab’s next journal club. Therefore, I’m going to check in on the code for the one we chose.

This month’s paper: Romain, S et al. Representation and detection of inversions in pangenome graphs. NAR Genomics and Bioinformatics 2026. doi: 10.1093/nargab/lqag125

Original code

The scripts for this paper are on GitHub.

Critique

Standard disclaimer: issues with published code are not necessarily anyone’s fault, and often are due to nothing more nefarious than time constraints.

Commit to comments

00_extract_multifasta_from_agc.py:

#
paths: list = list()
with open('temp.txt', 'r', encoding='utf-8') as reader:
    for line in reader:
        paths.append(line.strip()[:-2])

I’m not leaving anything out. That is just an empty comment.

This section is fairly self-explanatory as to what it’s doing (except that I’m unsure why the last two fields are being cut off), so I wouldn’t have been surprised if there was no comment. However, if there’s going to be a comment, leaving it empty does nothing but cause confusion. I pause and wonder if there’s invisible text or part of the code isn’t loading or similar.

Remove unused code

01_ragtag_chrom_contig_assoc.sh:

# Filter chromosomes to use from reference genomes
cd $OUTDIR
chrom_to_exclude="ragtag_exclude_extra.txt"
#grep '^>' $REFERENCE | awk '{ print $1 }' | sed -r 's/>//g' | grep -F '_' > $chrom_to_exclude
#echo 'chrEBV' >> $chrom_to_exclude
echo 'chrmitochondrion' > $chrom_to_exclude

From the looks of it, those two commented-out lines were previous things that were being used but are no longer. Leaving them in is odd. If there’s something useful there (e.g. maybe that logic should be used in some cases), then that should be explained, probably by another comment. Or if those are simply old versions that are no longer needed, then best would be to delete them, instead of having them hang around ghosting into the reader’s working memory.

Use spellcheck

At least on READMEs, maybe?

Script: 03_extract_select_contigs.sh Lists each haplotypes contigs conrresponding to a given chromosome, then extracts them using samtools faidx.

There shouldn’t be an “n” before the “r”s in “corresponding”.

I know that it’s not practical to run spellcheck on code files (the red squiggly lines for my C++ code would give me a conniption), and I know that I typo often. Still. READMEs are meant to be read by other people. One kindness you can do for those other people is making sure things are spelled correctly.

For example, I spellcheck these blog posts by copy-pasting a first draft into a browser tab (https://rot13.com/ to be specific, since it’s a perfect no-frills ad-free textbox) and letting the automatic spellcheck there do its magic. It’s very quick for me to glance through all the possible errors and dismiss false positives. I catch typos nearly every time I do this.

The common thread between all of these is that it can really pay off to do one final run-through of anything you provide to other people, just to wash off the artifacts which may have ceased to matter to you but would matter to them.

If there’s a recent paper you’d like me to look through, shoot me an email. Address in my CV.