The implications of alternative splicing in the ENCODE protein complement