To annotate as many proteins as possible, including predicted proteins,
each protein was mapped to its corresponding SwissProt protein (when available).
Thus, descriptions from individual proteins in each OrthoMCL cluster
were combined.
These yield to two annotations: