I can't reproduce the error, so I'm not sure how to help directly. But I highly suggest that you use the pipeline that uses a job scheduler. The
scripts and the pipeline
description are for
slurm, but the submission scripts should be easily adapted to other job schedulers. A big advantage of the pipeline that uses a job scheduler is how it handles missing data, in case some sequences are missing for some loci. For a given set of 4 taxa, the BUCKy analysis on these 4 taxa uses all the loci that have data for these 4 taxa, even if those loci may lack the sequences for some other taxa.
If some loci are missing a few taxa, those loci would be excluded entirely in the pipeline that does not need a scheduler.
Cecile.