In the last post in this series, we examined Casey Luskin’s attempt to minimize the level of identity that the human and chimpanzee genomes share. Luskin, with his concern to discredit the evidence that humans and chimpanzees share a common ancestor, suggested that the genuine identity value for these two genomes might be as low as 70%. Accordingly, we took the time to explain why his approach was misguided, and in the process learned about how genomes are sequenced and compared. As we saw, the human and chimpanzee genomes are in fact highly identical to each other, with identity values at 95% or more, depending on how the comparison is done. Such values are obviously highly consistent with shared ancestry, which would have humans and chimpanzees arriving at our differences by subtle modifications of our ancestral, shared genome.
Redundancy and codon usage
Having attempted to make his case for dissimilar genomes for humans and chimpanzees, Luskin turns to a second line of argument from my paper – evidence from the redundancy of the genetic code. This evidence now focuses on only a small portion of our genome – the sequences used to code for proteins. This portion of our genome is even more identical to chimpanzees than the average value – in fact, our protein-coding sequences are 99.4% identical at the DNA level.
The DNA code for making proteins is based on “reading” three DNA letters at a time to specify an amino acid. In this way, the protein-making machinery of the cell (a ribosome) translates what was once the DNA sequence of a gene three letters at a time to assemble a protein with the correct order of amino acids. These three letter groupings are called “codons”, and since there are four options (adenine, guanine, cytosine, and thymine) available for each of the three positions, there are 64 possible codons (4 x 4 x 4). Three of these codons – tag, taa, and tga – specify “stop adding amino acids to the protein,” leaving 61 to code for amino acids themselves. But since there are only 20 amino acids used in proteins, this means that the code has redundancy: most amino acids can be coded for by alternative codon options, and still produce the correct amino acid sequence.
Given that human and chimpanzees have 99.4% identical coding sequences, it should be easy to appreciate humans and chimpanzees, in the vast majority of cases, use the same codons in the same order even when alternatives exist. In my paper, I discuss the insulin gene in mammals as an example. We can use a short segment of this gene here to illustrate the pattern we observe.
Notice that the amino acid sequence (written underneath the DNA code) for this region of insulin is identical in all of these mammals, but the underlying DNA code, as you can see, has some differences. Humans, chimpanzees and gorillas are identical for this region; orangutans differ from the other primates by one base pair, and so on. Taking redundancy into consideration, there are many, many different ways to code for insulin with alternative DNA codes – for example, the segment above has two different codons for glycine (Gly) – the fourth codon in this segment is “ggg”, but the seventh, also a glycine, is “ggc”. Note how these two codons are identical in all the species listed here – even though either codon would code for a glycine (and indeed, so would “gga” or “ggt”). This pattern is not an isolated case – it applies to the entire gene, and, indeed, to genes in general for these species. The point is this: a designer could make these sequences much different at the DNA level, and still have the right amino acids to get the job done. Had a designer wished to avoid the impression of common ancestry for separately created organisms, it would have been as simple as selecting different sequences to code for the same amino acids (assuming that the designer “needed” to make the same proteins for both organisms in the first place).
Of course, once again we observe precisely the pattern one would predict common ancestry to produce. Primates are nearly identical because they more recently shared a common ancestor; primates and bats are more distantly related; finally, primates and mice are more distantly related still. Once again the face value of the pattern strongly suggests common ancestry – and as such, Luskin is at pains to find a non-evolutionary explanation for it, as best he can.
Function, function, everywhere!
The case that Luskin attempts to make is this: the reason the human and chimpanzee genomes are so identical is this: they need to be in order to function properly. For Luskin, those functional requirements extend all the way down to the codon options for each and every amino acid, in every gene in our genome, except in the rare cases where we see differences between us and chimpanzees.
If that seems a little extreme to you, I would agree – but recall, humans and chimpanzees are 99.4% identical for their coding sequences. If Luskin is going to argue that the reason for this identity is functional constraint, then that functional constraint needs to be pervasive.
Luskin attempts to build his case by citing papers that legitimately show cases where there is functional constraint at the DNA level for specific codon choices. The biology here is interesting, to be sure, but Luskin doesn’t come even close to establishing that the entire coding sequence of the human and chimpanzee genomes is under such constraint. The examples Luskin cites are cases where only a handful of codons are under constraint within a gene. For example, we know of cases where codon choice is important in order to determine the speed at which the ribosome does the translating. During translation amino acids connected to short nucleic acids called transfer RNAs (tRNAs) are brought in to recognize and bind to the codons. This means that for the different codons for a given amino acid, there are different tRNAs, and these different tRNAs may be present at different levels in the cell. If a particular tRNA is rare, then a codon that requires it will have to “wait” fractions of a second longer than if it was a codon with a common tRNA. This “wait,” in some cases, can be important for the protein to fold correctly. Other work has shown that codon choice can influence the binding of proteins required for the proper regulation of some genes – again, interesting work, but it applies only to a portion of some genes. Additionally, the constrained codons required for this protein-binding function are overwhelmingly right at the start of the genes in question – whereas we see 99.4% identity for the entire coding sequence, not just in these areas.
Even beyond these problems, there is an even greater problem for Luskin, and one that he does not address at all – human variation in coding sequences. Put briefly, we’ve known since the early/mid 2000’s that humans have a significant amount of variation not only in non-coding sequences, but also in our coding sequences. This variation, documented by the HapMap project, shows that the absolute functional constraint that Luskin needs to make his argument work simply is not to be had – humans can and do vary significantly in their coding sequences, and a large amount of this variation is in the form of alternative codons for a given amino acid. If humans can vary in their coding sequences without functional problems, then we have empirical evidence that these sequences are not functionally constrained in the way that Luskin requires. Additionally, now that we have been sequencing multiple chimpanzee genomes, we see the same sort of variation in chimpanzees – further exacerbating the problem for Luskin.
So, the only way for Luskin’s argument to work is if there is functional constraint at virtually all codons in the human and chimpanzee genomes, and that these functional constraints, for reasons he does not discuss, need to be almost identical for both species. The first claim (as we have seen) is demonstrably wrong: there is significant variation within human and chimpanzee populations for coding sequences, showing that absolute constraint is absent. The second, implied claim, is curious: would not a designer, creating the human and chimpanzee genomes from scratch, be free to constrain the function of the two genomes in different ways? Protein/DNA binding patterns, translation speeds based on tRNA pools, and other such constraints (even for the small portion of the coding sequence they constrain) need not be constrained in (almost) exactly the same way in two separately created organisms, surely? If so, then I suppose Luskin’s argument is that the designer was somehow unable to avoid the (erroneous) impression of common ancestry for humans and chimpanzees because the designer was unable to engineer them in a distinct manner due to functional constraints (despite the observation that all other mammals were successfully engineered to be less similar).
It seems instead that the most reasonable conclusion from the evidence is that the Creator brought about the diversity of life – including humans – through common ancestry.
In the next post in this series, we’ll tackle Luskin’s arguments about the conservation of gene order on chromosomes: the genomic evidence for common ancestry from synteny.