Think globally and solve locally: secondary memory-based network learning for automated multi-species function prediction

Think globally and solve locally: secondary memory-based network learning for automated multi-species function prediction