The protein families we know of are based on a very small number of full genomes. Even if we have conservative assumptions - i.e. that new protein families are

The protein families we know of are based on a very small number of full genomes. Even if we have conservative assumptions - i.e. that new protein families are not uniformly distributed, but obey some kind of power law, where the first few sequenced organisms will contain most of the families; given how few species are sequenced, I expect that there are many more families. The title would be better if it said "protein family" or "protein class". New transcription factors are discovered fairly often. Humans have around ~2K transcription factors. New families, less often.