Paula Zahn

• Jason Lisle [Ph.D. in astrophysics, Cincinnati, affiliated with a pro-creationism group called Answers in Genesis] — Well, there's no neutral ground, is there? I mean, you're ultimately either for what God has said as word or against it. And that's what the real issue is here.
• Paula Zahn — Eugenie?
• Eugenie Scott [director of the National Center for Science Education, San Francisco] — No, we're treating this as if there are two alternatives, evolution, and the Institute, or the Answers in Genesis' version of creation.

Among all of the many examples of mobile elements or "parasitic sequences" that affect the function of the human genome, this paper describes several examples of functioning genes whose sequences have been almost completely derived from mobile elements. There are many examples where the synthetic coding sequences of observed mRNA sequences are made up of mobile element sequences, to an extent of 80% or more of the length of the coding sequences. In the examples described here, the genes have named functions, and some of these functions have been studied. It appears that each of the functioning genes was originally formed from mobile elements and that in some process of molecular evolution a coding sequence was derived that could be translated into a protein that is of some importance to human biology. In one case (AD7C), the coding sequence is 99% made up of a cluster of Alu sequences. In another example, the gene BNIP3 coding sequence is 97% made up of sequences from an apparent human endogenous retrovirus. The Syncytin gene coding sequence appears to be made from an endogenous retrovirus envelope gene.

Roy J. Britten, "Coding sequences of functioning human genes derived entirely from mobile element sequences" [

In our theory, duplication and divergence under positive selection can adapt and optimize a gene within a narrow range only. If abetted by directed mutation, as many examples imply, this process is quite remarkable. However, to go outside their narrow ranges and acquire new genetic programs, genes require new sequences that pointwise mutations will never compose, we think. If instead, these sequences are acquired by gene transfer, then powerful genetic operating systems with accessories like mobile elements, and processes like recombination and DNA repair, may subsequently piece them into place. If we're right, any diverged genes comprising wholly new genetic programs will exhibit whole insertions, and possibly deletions, of exons or similar-sized sequences. IHGSC's current analysis specifically ignored likely insertions and deletions in the studied duplicates, and ignored other possible duplicates altogether, if likely insertions and deletions were longer than the remaining sequence. Unsurprisingly, all of the duplicated genes in this study appear to exhibit adaptation and optimization within narrow ranges. None exemplifies the birth of new a function.

Can a diverged gene acquire a new function by pointwise mutations only, or must it undergo exon-sized recombinations? We suspect the latter. We see nothing in this study that indicates otherwise.

ragworm eyes

Not everyone is convinced of this scenario. The Germans themselves "detected a highly conserved stretch of amino acids... that is shared among the [vertebrate] c-opsins but distinct in the [invertebrate] r-opsins." A connection between the c-opsins in ragworms and vertebrates is much better supported. Sequence similarities suggest that an evolutionary path between them could have been traversed by pointwise mutations.

A friend asked us if this research illustrates the darwinian evolution of eyes. We think instead that it shows how essential, conserved programs may adapt and be optimized within narrow ranges. And even if, as hypothesized, c- and r-opsins evolved from a single predecessor, it was probably also, already an opsin.

Compared to its predecessor, the enzyme encoded by GLUD2 has 18 amino acid substitutions. The two that were fixed earliest are 100% conserved and are essential for the newer enzyme's brain-specific function. Another nine substitutions characterize differences among different apes and humans. The remaning seven aparently have little or no affect on function. Meanwhile, the enzyme encoded by GLUD1, lacking brain-specificity, has apparently remained unchanged. Both enzymes are versions of glutamate dehydrogenase (GDH).

We agree that gene duplication and divergence occur, and we have already supposed that intronless paralogs can be or become functional. Furthermore, we suspect that directed mutation — accelerated sequence changes in the right places — may assist "positive" selection. But we believe that these processes can only optimize existing genetic programs within very narrow ranges; we doubt that they can write new programs. The enzymes encoded by GLUD2 and GLUD1 differ by only a few amino acids, and both enzymes are versions of GDH. To us, this is a story about optimization within a narrow range. We think Ohno's theory still lacks confirmation.