fix asterisks in gene names

[GWMcElfresh]

This is a small bugfix toFitRegularizedClassificationGlm() to allow specifically gene UGT2B9*2 (the asterisk being the problem) to pass through MLR3's pipelines.

[bbimber]

@GWMcElfresh: just one comment. You are doing things like this:

gsub("\\*", "star", colnames(target_labeled_data))

this would break if a gene happened to contain the string star, right? same for dash, etc. Could we do anything to make the replacement strings more likely to be unique? Can we use any other characters, such as:

star
⭐
-star-

or anything that would make these placeholder strings more likely to never occur in a real gene name?

[GWMcElfresh]

@GWMcElfresh: just one comment. You are doing things like this:

gsub("\\*", "star", colnames(target_labeled_data))

this would break if a gene happened to contain the string star, right? same for dash, etc. Could we do anything to make the replacement strings more likely to be unique? Can we use any other characters, such as:

star ⭐ -star-

or anything that would make these placeholder strings more likely to never occur in a real gene name?

I checked out lower case letters in the mmul10 genome and the only lower case occurrences are ORF (e.g. "C15H9orf24", "C9orf72"), so I think we're safe here so far.