UmiAllA #

[[filter]]
    mode = 'umi_all_a'
    action = 'remove|keep_only' # see 'Filter steps' for details

Filter reads that have all ‘A’ bases in their UMI.

Why is this here? #

This is what STARsolo (2.7.11b) does.

Mistakenly.

The doc and code suggests STAR should filter all homopolymers but there’s a bug where the UMI length is not read correctly from --soloUMIposition, and that makes all homopolymer checks but A always fail, except if your UMI length happens to be 10, or you’re setting --soloUMIlen in addition.

(It’s claimed to be fixed in STAR 2.7.11a, but I’ve observed it even later).

You probably do not want to use this in production., but UMI holopolymer instead.