The Talk.Origins Archive Post of the Month: March 2006

Subject:    | Pseudogenes
Date:       | 02 Mar 2008
Message-ID: | 811b99dc-976e-45a1-8ae4-0492c3aaff15@8g2000hse.googlegroups.com

by Sean Pitman:
>>> At first you claimed that pseudogenes were non-functional garbage "junk"
>>> DNA.

by Howard Hershey:
>> 'Most' pseudogenes probably are initially or wind up being
>> 'non-functional' garbage. The ones that are *preserved* (and thus
>> seen in current genomes) are not *most* pseudogenes. They are a
>> subset of pseudogenes.

by Sean Pitman:
> You are quite hilarious sometimes Howard. Where do you think the term
> pseudogene came from? Let me tell you. It was a term first used to
> describe certain genetic sequences that looked sort of like genes, but where
> thought to be damaged so that they couldn't make proteins any more and were
> therefore thought, originally, to be non-functional.

I will be officially agnostic on whether 'most' pseudogenes, that is genes which have what *appears* to be a coding sequence that arose by simple classical duplication or by insertion of a processed reverse transcript but which has stop codons or frameshifts that *do* prevent, at minimum, the translation of a full-length duplicate protein like that which the modified duplicate appears to be from.

My personal opinion is that 'most' processed pseudogenes, which often are not transcribed in either direction (transcription of an anti-sense RNA can have a biological effect) as well as not translated and which lack the regulatory DNA sequences that the original gene sequence had will be as a 'functionless' sequence. Such sequences, being selectively neutral will accumulate changes as a function of time since their 'creation' as an insert. This assumes that the site of insertion does not itself generate a 'mutational' effect unrelated to the sequence of the insert. An insertion in a coding sequence can certainly have a selective effect (again, that effect can be beneficial, detrimental, or selectively neutral) on that gene.

*If* a processed pseudogene does have its antisense strand transcribed, this can potentially have a 'biological effect'. Any such biological effect can be either beneficial, detrimental, or selectively neutral. In *most* cases, the biological effect of a processed pseudogene that arises naturally will be selectively neutral. However, some may *initially* be detrimental because of the consequences of the antisense RNA in reducing the *level* of translation of the *real* sequence. If this happens, merely more mutations in the pseudogene sequence will not, until many more have accumulated, lead to a reversal of the dampening effect of the antisense RNA. Interestingly enough, one method by which a mutation can lead to a correction of this dosage-effect would be an upregulation of transcription from the original gene leading to dosage compensation. That could easily lead to a state of "irreducible complexity" in which the originally deleterious pseudogenes depression of translation has become necessary to prevent overtranscription from the original gene!

But that would be a relatively rare consequence of processed pseudogene insertion. Most processed pseudogenes will be selectively neutral in their consequences and we will see a time-ordered pattern of loss of identity with the original source material. That is, old processed pseudogenes will have accumulated more differences from the original source material (which will accumulate selectively neutral changes itself) than will young processed pseudogenes. Only rarely will there be significant conservation of pseudogene sequences over time.

Being a *possible* future source of coding sequences is not a *function* of the sequence in the present time because there is no 'goal-oriented' or 'teleologic' mechanism that can preserve a sequence for a potential future use. Natural selection, which is the *only* mechanism that can preserve sequence from neutral drift, does not work on 'potential future utility'. It only works on current utility.

Pseudogenes that are direct duplicates have a greater capacity for having some biologically relevant consequences. Duplicate genes are, of course, the major engine of evolution of new function (although not intentionally so). But when they accumulate a stop codon or a frameshift so that they cannot make the *full-length* translational product that the *original* source produced, they become, by definition, a pseudogene.

But, depending on the sites of mutation, a duplicate that cannot produce a full-length translation product still can produce a partial translational product, a full or partial transcript, a reverse transcript, and it could still have functional regulatory sequences. All of these *could*, but need not, have *some* biological activity precisely because the original source gene has those biological activities. Sometimes these biological activities can be beneficial, sometimes detrimental, and sometimes selectively neutral.

*When* such a function is beneficial, selection will preserve it. *When* such a function is deleterious, selection will cause mutations that compensate for the deleterious effect, which usually is due to some form of difficulty with dosage compensation. Sometimes these compensating mutations will lead to "irreducible complexity", like that described above for processed pseudogenes, where the compensating mutation makes the originally deleterious duplication now a sequence needed to produce the right amount of real product.

Sean's claim is that *most* pseudogenes are actually *designed* intelligently in order to perform desirable or necessary function. I don't think he actually has evidence that *most* pseudogenes that form actually are necessary. That would require more evidence than he has presented, especially since there would be a bias for the preservation of those parts of a pseudogene that do have functional utility, even if that functional utility is nothing but a method for compensating for the deleterious features of the original duplication event.

I could tell you how one could go about getting an unbiased analysis of pseudogene activity, but I don't think Sean is really interested in that. And I, frankly, don't care what the answer would be. I don't have any problem with being wrong when I say that *most* pseudogenes do not serve any important biological function.

Sean, because he is claiming that each and every pseudogene *must* be designed intentionally is the one with the problem. It is not sufficient for him to claim that *most* pseudogenes have *some* biological function. He must show that *all* pseudogenes have a currently useful function (being present for possible future use is not a current function) and that *all* parts of the pseudogene must have the particular sequence they do to perform the particular function that pseudogene has in his design scheme.

What Sean has to explain is *why* an intelligent designer would choose to produce those functions that pseudogenes do have (remember that pseudogenes, by definition, cannot produce the full-length translation products implied by their sequence) by generating specific pseudogene sequences that are not able to be translated. Sean has to explain why the functional pseudogenes have to have the exact pattern of mutations that would prevent full-length translation product synthesis that they do have in order to perform this function. Why would a designer *intentionally* design such systems to look exactly like what would happen via an observable type of mutational error (duplication or reverse transcript insertion) followed by random mutation producing a loss of function mutation for the pseudogene? And Sean needs to explain why the mutations in these functional pseudogenes exist in a pattern consistent with common descent and with the varying appearance of greater degradation of sequence in many pseudogenes that would be consistent with neutral drift in the coding sequence over time. IOW, Sean needs to explain why his designer.

I have no problem with a duplicate protein having some biological activity. I have no problem with pseudogenes having certain types of biological activity. I suspect, that like most mutations, most pseudogenes, which we know can arise by naturally occurring mutations, will either be selectively neutral or selectively detrimental (mostly due to dosage effects). Some duplicates will be immediately beneficial. All organisms with duplicates can undergo further mutation to modify the duplicates to specialized functions, to compensate for deleterious effects, to retain duplicates that have beneficial effects, to generate modified functionality of current systems. That is, duplications or processed duplications are mutational events. Selection will determine the fate of these events.

[snip diatribe, which does not support his view that *most* pseudogenes are intentionally designed to look just like defective duplicates of originally fully translated proteins.]

Are (non-functional) pseudogenes a product of Intelligent Design?

Post of the Month: March 2006