scNapBar fails because of the random shuffling of barcodes

[eboileau]

In rule build_genome, a fraction of numIlmReads is used to simulate reads we want to use as true negatives. For this we shuffle the bk_barcodes.tsv.gz. To facilitate reproducibility of the results, I added a seed to shuf using the very file bk_barcodes.tsv.gz, so that it will "randomise" in the same way every single time. The problem is that it needs a larger number of bytes, otherwise an end of file error is thrown. If dealing with e.g. Visium Spatial 10X, the number of total/raw barcodes is much lower, and fixed to 4992, which is the number of spots in the spatial array, so this results in an error, for large numIlmReads (for a given percent_raw).

This is easily fixed by "seeding" using the Illumina BAM file, which should in most cases be sufficient.