Support for semicolons to separate proteins

[MatthewThe]

The triqler input currently expects all proteins mapping to a peptide to be separated by tabs.
This was originally done to mirror the percolator output file format.
However, this does not play well with popular Python packages such as pandas and many proteomics search engines already use a semicolon to separate protein identifiers.

Allowing the user to specify a custom protein separator (as Percolator introduced in v3.07) would be a good idea.