Tabix requires a header to be specified with a special character (default '#'). One can create a tabix-indexed file without it by skipping the header row with e.g -S1, but then there will be no header row.
Autoreporting unfortunately uses pysam.TabixFile.header-member variable, which is empty for those files... So things crash.
Mitigations:
Instead of using pytabix, fetch the header using e.g. gzip.open(). Then we open up the space of allowed files to be tsvs with a single header row, which is much saner than "files that are tabix-indexed, but also have to have a tabix-compatible header row". This also makes it easier to run the pipeline for e.g. just annotation & top report output.
Lipastomies
commented
1 year ago
Tabix requires a header to be specified with a special character (default '#'). One can create a tabix-indexed file without it by skipping the header row with e.g -S1, but then there will be no header row. Autoreporting unfortunately uses
pysam.TabixFile.header-member variable, which is empty for those files... So things crash.Mitigations: