Updated project metadata. Computational approach to build non-redundant protein databases for strains with unknown genomic background