Well that was a pain. Trying to get recount3 unify pieces set up for zebrafish.

1 - Current primary assembly has 993 pieces, then add 99 from ERCC & SIRV. Docker image of `perbase` crashes on chromosome 1032.
2 - OK, filter to the primary 25 using `seqtk`.
- Default is to put each chromosome on one line.
- Now `perbase` crashes when it reads in the first chromosome.
- Learn how to have `seqtk` wrap lines at 60 characters.

It finally works! Well, I haven't gotten it read into R, but ...