Thursday, 5 March 2015

Ron Cowie's response to the DNA evidence for evolution

The recently-launched Christadelphian Answers website is a positive development in our community as it correctly recognises the considerable challenge the "New Atheism" movement poses for our community. However, its effectiveness is compromised by poorly researched attacks on evolutionary biology. The danger here is that potential converts will dismiss our message concluding that if we make basic mistakes in attacking evolutionary biology, we will not be trusted to correctly expound the Biblical message of salvation.

In "What About the DNA evidence for evolution", the author, Ron Cowie has badly misunderstood why biologists regard DNA similarity as evidence for evolution:
It is argued that DNA similarity between species is consistent with these species having a common ancestor. This scenario is true of siblings within a family –  they share 50% of their DNA and have the same parents. But is it necessarily true that all organisms sharing common sequences of DNA must have a same ancestor? 
DNA forms the blue print of life and is a chemical code describing the building blocks of life and how they are to be assembled. Species living in similar environments, with similar biological functions and made from similar building blocks are likely to have similar DNA sequences. Similarities in DNA sequences are simply measures of similarity between species. Concluding that a common ancestor is required is not logically necessary, and is an interpretation of an observation rather than a proof of evolution. [1]
Cowie has not grasped the reason why evolutionary biologists regard sequence similarity as powerful evidence for common ancestry. A quick example will make this clear.

Proteins are composed of small building blocks called amino acids. Each amino acid is encoded by a three nucleotide sequence called a codon. As there are on average three possible ways to code for each amino acid, for a moderately long protein such as cytochrome C (critical in energy metabolism) which consists of around 104 amino acids, there are 3^104 or around 4x10^49 ways to code for the exact same protein, a phenomenally large number.

"RNA-codons" by TransControl - Licensed under CC BY-SA 3.0 via Wikimedia Commons.

Given the redundancy of the code, it is possible for a gene to pick up neutral mutations which change the gene, but not the protein sequence for which it codes. If common descent was true, then we would predict that species that share a recent common ancestor would have gene sequences that differ by fewer neutral mutations, while those with remote common ancestors would differ by many neutral mutations. Conversely, there is no reason under a model of special creation for us to expect this pattern. In fact, the possible gene space is so large that every species could have its own cytochrome C gene sequence differing from every other species by so many point mutations as to invalidate common descent.

What we see confirms common descent. The family trees constructed from gene sequences confirms the predicted evolutionary family tree. As evolutionary biologist Douglas Theobald notes, with respect to cytochrome C:
"...the cytochrome c proteins in chimps and humans are exactly identical. The clincher is that the two DNA sequences that code for cytochrome c in humans and chimps differ by only four nucleotides (a 1.2% difference), even though there are 1049 different sequences that could code for this protein. 
"The combined effects of DNA coding redundancy and protein sequence redundancy make DNA sequence comparisons doubly redundant; DNA sequences of ubiquitous proteins are completely uncorrelated with phenotypic differences between species, but they are strongly causally correlated with heredity. This is why DNA sequence phylogenies are considered so robust." [2]
The fact that humans and chimpanzees DNA is nearly identical is something that greatly troubles special creationists, and they have attempted to dismiss this evidence by claiming, as Cowie does, that this depends "on which type of genetic variation is being measured". Although Cowie does not mention him, other special creationists cite an article by special creationist Jeffrey Tomkins in which he claims that the similarity is only 70%. His figure, as neurologist Steve Novella points out, is quite misleading:
What about the percent? With this type of analysis scientists come up with a range, about 95-98% similarity between human and chimp. The average number of 96%, derived after the complete mapping of the chimp genome, is now generally accepted. But this number is just representative. It depends on what you’re looking at. Some genes are highly conserved, others less so. 
So how does Tomkins come up with 70%. Well, he is not comparing point mutations of aligned segments. He is comparing chromosomes to see how many segments line up to some arbitrary amount. As many others have already pointed out, this result is not wrong, it’s just irrelevant. Well, it might also be wrong. Others have found it difficult to reproduce his results. But even if his analysis is accurate, it is simply the wrong analysis to apply to dating the last common ancestor. 
To explain the problem further, he is applying mutation rates for point mutations (changing a single base pair) to other types of mutations, like gene duplications or insertions, that might change thousands or millions of base pairs with a single mutation. He is essentially treating a single mutation that results in the insertion of 10,000 base pairs into the genome as if it were 10,000 separate mutations of single base pairs.
What completely undercuts the special creationist argument is that when we compare human, chimpanzee, and monkey DNA, we see a greater distance between human and monkey than we do with human and chimpanzee. This is exactly what common descent would predict, but is inexplicable under a model of special creation. Biochemist Larry Moran makes this point elegantly in a recent blog post that referenced a 2009 paper which compared genetic variation within human genomes, between human and chimpanzee genomes, and between human and macaque genomes:
Notice that the substitutions are pretty much randomly scattered over every part of the two chromosomes. The data is consistent with the idea that most of the DNA in those chromosomes is junk and most of the substitutions are nearly neutral mutations fixed by random genetic drift. The differences between each pair of species is consistent with an approximate molecular clock corresponding to a constant mutation rate over million of years. The absolute levels of sequence identity (i.e. 98-99% for chimp/human) is consistent with the time of divergence from a common ancestor based on the fossil record and other criteria.  
Here are my questions. Is there any other explanation that accounts for the data? Is it possible to explain the results as adaptations—substitutions that are mostly fixed by natural selection? Is it possible to explain the results according to Intelligent Design Creationism?

(Human-Human variation. Human-Chimp Variation. Human-Macaque variation)

Needless to say, there is no credible special creationist answer, other than to argue that God made things this way, which is merely a way of failing to answer the question.

Cowie's rebuttal of shared 'mistakes' in DNA ignores the reason why it is regarded as overwhelming evidence for common descent:
"In the past, DNA was only considered functional if it was translated into protein. This meant that the 98.5% of DNA, which is not translated into protein, was called “junk DNA”. It is now thought that 80% of the genome has a function, even if it is not translated into proteins. There are also suggestions of an advantage for humans in not producing their own Vitamin C internally, but relying on their food as the source. This example of a DNA mistake may not be a mistake after all. Human knowledge is incomplete and it is a mistake to conclude that no known purpose is the same as no purpose at all."
Cowie commits many errors in this paragraph. While only a small portion of the genome directly codes for proteins, it is incorrect to assert that biologists dismissed it as 'junk'. T.R. Gregory, an expert on genome size evolution and 'junk DNA' points out [3] that introns, the large spaces of DNA in genes that are spliced out of the RNA copy of a gene and do not contribute directly to protein coding, were believed by scientists to have some function, and cites the following:
“Perhaps the most surprising discovery in the initial studies of eukaryotic gene structure has been that many genes contain interruptions in the coding sequences. The origin and the function of these intervening sequences (IVS or introns) are not yet well understood but are the subject of intense investigation.” - Wallace, R.B., P.F. Johnson, S. Tanaka, M. Schöld, K. Itakura, and J. Abelson. Directed deletion of a yeast transfer RNA intervening sequence. Science (1980) 209:1396-1400. 
“Since the discovery that many eukaryotic genes are discontinuous, a number of studies have been directed towards identifying a function for intervening sequences (IVSs).” - Johnson, P.F. and J. Abelson. The yeast tRNA(tyr) gene intron is essential for correct modification of its tRNA product. Nature (1983) 302: 681-687. 
“It is possible that the relationship between the location of the splice junction in the gene at the surface of the protein confers a biological advantage and hence is a result of natural selection. Introns and their associated splicing systems could be exploited in many ways during the evolution of a protein.” - Craik, C.S., S. Sprang, R. Fletterick, and W.J. Rutter. Intron-exon splice junctions map at protein surfaces. Nature (1982) 299:180-182. 
“We conclude from this experiment that the intron in the yeast actin gene does not have an observable function. It is possible that the role of the intron is too subtle to be observed in laboratory conditions of growth or that the intron, while having evolutionary significance, has no present role. To conclude that this is true for all yeast genes that contain introns would of course be premature, but there exist strains in which mitochondrial introns have been removed with no observable effect.” - Ng, R., H. Domdey, G. Larson, J.J. Rossi, and J. Abelson. A test for intron function in the yeast actin gene. Nature (1985) 314: 183-184.
Over the intervening 30 years, we have found functions for some of this non-coding DNA, but have also found that much of it - at least 66% - has no function. Around 45% of our genome consists of retrotransposons, mobile genetic parasites that copy and paste themselves randomly, with another 8% consisting of decayed remnants of ancient viral infection. Apart from the tiny percentage that has been secondarily co-opted to perform new functions, most of this is functionless junk.

Cowie unfortunately cites the grossly hyped ENCODE papers, as well as the claim that 80% of the genome is functional. This is incorrect. Ignoring the fact that functional does not mean essential (an active retrotransposon copying itself over a working gene is functional but not beneficial). ENCODE head scientist Ewan Birney, in a BBC interview acknowledged that around 10-20% of the genome is vital for life, [4] a far cry from the 80% figure.

Cowie simply does not address the reason why the GULO data (not to mention the hundreds of other shared identical pseudogenes, remnants of ancient viral infection, retrotransposons and other genetic markers) is regarded as compelling evidence for common descent. The reason is simple, logical, and compelling. Humans, apes and monkeys cannot make their own vitamin C because the gene that codes for the final enzyme in the biosynthetic pathway is broken. Furthermore, it is broken in exactly the same way, and the pattern of point mutations the pseudogene has picked up is exactly what we would expect if common descent was true.

Cowie claims that there are advantages for humans in not producing their own vitamin C internally, and cites two papers as evidence. He has again missed the point. Whether the GULO mutation conveyed a selective advantage is independent from the question of its evidence for common ancestry. Evolutionary biologists would have no problems with the original GULO mutation occurring in the common ancestor of humans, apes, and monkeys and being positively selected if its net benefits exceeded its net drawbacks, much in the same way that in malaria endemic countries, the mutation for sickle cell anaemia has been positively selected as its net benefits outweigh its drawbacks. [5] A beneficial mutation nonetheless has a natural history and can be used to show common ancestry.

This is the point ironically made by one of the papers Cowie cites:
Was loss of this enzyme just a random mutation in animals, having no disadvantage because these species were consuming Vitamin C in the diet? Or did its loss perhaps have a selective advantage, since gulonolactone oxidase is an H2O2-producing enzyme [6]?
It is odd that Cowie cites this paper given that it does not argue against evolution. The simple fact is that outside of a tiny fundamentalist fringe that privileges human interpretation of the creation narratives over the witness of creation, the fact of common descent is entirely unremarkable. Even the intelligent design advocate Michael Behe accepts that the GULO data confirms human-ape common ancestry:
When two lineages share what appears to be an arbitrary genetic accident, the case for common descent becomes compelling, just as the case for plagiarism becomes overpowering when one writer makes the same unusual misspellings of another, within a copy of the same words. That sort of evidence is seen in the genomes of humans and chimpanzees. For examples, both humans and chimps have a broken copy of a gene that in other mammals helps make vitamin C As a result, neither humans nor chimps can make their own vitamin C. 

The same mistakes in the same gene in the same positions of both human and chimp DNA. If a common ancestor first sustained the mutational mistakes and subsequently gave rise to these two modern species, that would very readily account for both why both species have them how. It's hard to imagine how there could be stronger evidence for common ancestry of chimps and humans. (Emphasis mine) [7]
Cowie's appeal to incomplete knowledge in order to imply that these genomic glitches may have function again overlooks the fact that it is the presence of identical genetic glitches at identical places in human and ape genomes, irrespective of secondary cooption of cuntion that is the question. Former human genome project head and medical geneticist Francis Collins puts it well:
Even more compelling evidence for a common ancestor comes from the study of what are known as ancient repetitive elements (AREs). These arise from "jumping genes," which are capable of copying and inserting themselves in various other locations in the genome, usually without any functional consequences. Mammalian genomes are littered with such AREs, with roughly 45 percent of the human genome made up of such genetic flotsam and jetsam. When one aligns sections of the human and mouse genomes, anchored by the appearance of gene counterparts that occur in the same order, one can usually also identify AREs in approximately the same location in these two genomes.

Some of these may have been lost in one species or the other, but many of them remain in a position that is most consistent with their having arrived in the genome of a common mammalian ancestor, and having been carried along ever since. Of course, some might argue that these are actually functional elements placed there by the Creator for a good reason, and our discounting of them as "junk DNA" just betrays our current level of ignorance. And indeed, some small fraction of them may play important regulatory roles. But certain examples severely strain the credulity of that explanation. 
The process of transposition often damages the jumping gene. There are AREs throughout the human and mouse genomes that were truncated when they landed, removing any possibility of their functioning. In many instances, one can identify a decapitated and utterly defunct ARE in parallel positions in the human and the mouse genome.

Unless one is willing to take the position that God has placed these decapitated AREs in these precise positions to confuse and mislead us, the conclusion of a common ancestor for humans and mice is virtually inescapable. This kind of recent genome data thus presents an overwhelming challenge to those who hold to the idea that all species were created ex nihilo. (Emphasis mine) [8]

Cowie's assertion that the DNA evidence for common ancestry is not convincing, has been falsified. There is arguably no stronger line of evidence for human-ape common ancestry than from comparative genomics.

Further reading

1. Cowie J “What about the DNA evidence for evolution?Christadelphian Answers Nov 18 2014 

2. Theobald, Douglas L. "29+ Evidences for Macroevolution: The Scientific Case for Common Descent." The Talk.Origins Archive. Vers. 2.89. 2012. Web. 24 Nov. 2014 <>

3. Gregory T.R. "Were introns immediately dismissed as useless junk?" Genomicron May 12 2014 

4. Gregory T.R. "BBC interview with Ewan Birney" Genomicron April 1 2013 

5. Sargent G.R. "The Natural History of Sickle Cell Disease" Cold Spring Harb Perspect Med 2013; doi:10.1101/cshperspect.a011783 

6. Halliwell B "Vitamin C and genomic stability" Mutation Research (2001) 475:29-35

7. Behe M “The Edge of Evolution. The Search for the Limits of Darwinism” (2007, Free Press) pp 71-72

8.Collins FS The Language of God: A Scientist Presents Evidence for Belief (2006: Free Press) p  135-137