Well that don’t look right at all

So they’re -9 compressed bz2 files

And when looking for the bzip2 header that indicates compression and start of file we see:

the first 8 lines are expeted, BZh and then the compression level wouldn’t be in 1-8 which were compressed with the associated compression levels

But the last two, uhhh, how did you possibly generate bzip2 files with that many headers? Apparently that can happen through concatenation.

Fun fact: bzip2 reads _2 fine.
Funner fact: basically no other implementations do. I.e. most bioinformatics tools. They just read the first entry and are done. But we only know this because it’s split mid-read, unlike _1 which runs successfully while actually failing.

FastQC reports 1927 reads which is off by, a lot. (451782 is the correct value.) We’d never know unless we carefully check this.

So if your tool breaks on a bzip2 file, try decompressing and recompressing, and updating your resume on linkedin while you find a new career.

#bioinformatics #software