Hi Chris,
I was able to research the matter a little bit more. A later/better study–by the same team that published the ENCODE study–shows that the human genome has 14,112 pseudogenes. Of these, only 876 (6.2%) are transcribed.
93.8% of pseudogenes in the human genome are not even transcribed. If they are not transcribed, how can they possibly be functional?
Now let’s examine those few pseudogenes that are transcribed. Xu and Zhang examined 34 such pseudogenes in the human genome and determined that only 5 (14.7%) have any functionality as evidenced by selection pressure.
Assuming that all of the 876 transcribed pseudogenes in the human genome have the same ratio of functionality (14.7%) as those examined by Xu and Zhang, we can infer that 129 of them are functional.
Based on this analysis, what proportion of human pseudogenes are actually functional? 129 / 14112 = 0.0091 – less than 1%.
Certainly there is some imprecision in this analysis. The proportion might be a little lower, or it might even be two or three times as large. There is no particular reason to think so, but scientific analysis being what it is, I would be willing to admit that possibility.
In any case, it is impossible for the percentage of functional pseudogenes to exceed 6.2% because that is the limit imposed by their transcription rate.
Thanks,
Chris Falter