it would be safer to explicitly require the user to list on the command-line which rule set to use

[UnixJunkie]

I feel, hiding away in a configuration file which rule set to use is a little dangerous, because this is indirect (some users will not know/realize which rule set is being used). For scientific software, I think things should be very explicit.

If you don't want to change the default behavior (because some users might already be used to it), then having command-line options to explicitly specify which rule set to use would be nice. E.g. --PAINS, --Dundee, etc.

[PatWalters]

Good idea, I'll add this