Changes:

    - Uses pyproject.toml instead of setup.py for installation
    - update minimal requirement for ncbi_genome_download to be compatible with the last NCBI format (issue #42)
    - use --amino option for mafft protein alignements in order to force sequence type to be proteic and deal with ambiguous aa (see https://github.com/gem-pasteur/PanACoTA/commit/728445dd84f8b5d159c97371ee828ea8761ea938#commitcomment-122306169)
    - fix issue with protein names. They must fulfill the conditions described in https://aperrin.pages.pasteur.fr/pipeline_annotation/html-doc/usage.html\#protein-files

Solves issue #42