A recent paper published by the Answers Research Journal, the research publication of Answers in Genesis, reported a comparison of human and chimpanzee genomes and found that they have, on average, a DNA similarity of only 70%. This is a very striking number since the usual numbers you hear thrown about as representing the similarity of the human and chimpanzee genome are usually 98.6%, 98%, 96%, 94% and even 90%. The variation in these values is partly due to larger data sets becoming available for comparison but is mostly due to different assumptions used to calculate the percentages. For example, the higher similarities are derived by examining only DNA sequences that correspond to actual pieces of genetic code (ie. the genes in the genome). The estimates of lower similarity reflect comparisons that include vast stretches of DNA that don’t include coding regions or are the spacers between genes. Now comes word from a paper published by Answers in Genesis that the average similarity of the DNA sequence is actually only 70%. What is up with that? Is this as significant a finding as it sounds? (see footnote)
First, let me say I am really not interested in entering into a protracted discussion about human origins. I not saying that humans aren’t a unique creation. Rather I am commenting on this paper because I think that it promotes a very poor scientific argument for the uniqueness of man and thus is one that Creationists should stop using. Tompkin and others at ICR and AIG have striven for many years to demonstrate that chimpanzees and humans could not be related because they are far too different, both genetically and morphologically, for anyone to reasonably believe they shared a common ancestor. They have attempted to build a genetic case against evolution and promote these differences as proofs that give the lay Christian confidence that there is no connection between humans and chimpanzees.
In this brief article I will argue that rather than providing strong case against primate-human ancestry, Tomkins and others actually undermine their own case for the uniqueness of man by their inconsistent approach to applying genetic distance as a principle of species (or “kinds”) delineation. Let’s start by looking at the conclusion of Tomkin’s article from Feb. 20, 2013: “Comprehensive Analysis of Chimpanzee and Human Chromosomes Reveals Average DNA Similarity of 70%.”
“Genome-wide, only 70% of the chimpanzee DNA was similar to human under the most optimal sequence-slice conditions. While, chimpanzees and humans share many localized protein-coding regions of high similarity, the overall extreme discontinuity between the two genomes defies evolutionary timescales and dogmatic presuppositions about a common ancestor.”
Tomkins makes it abundantly clear that he believes that such extreme genetic dissimilarity between humans and chimps should make it ridiculous for anyone to think that they humans and chimpanzees could have a common ancestor. How does he come to this 70% number when other geneticists derive much higher values. First, Tomkins states quotes a “geneticist” named Buggs that apparently had already calculated a 76% similarity between the genomes. The implication seems to be that scientists already knew that there was a big difference but have been ignoring it. However, I tried to follow the link in the references to this Buggs quote but the link (only 7 days after publication!) is already dead or was printed in error. (see comment below for reader supplied links). Regarding how Tomkins derives the 70% value, I really am not certain because the methods described in the paper do not allow me to fully ascertain how the analyses was done. However, Tomkins clearly is comparing all parts of the genome and depending on how they align is counting not only differences in the sequences, but gaps where there are insertions and deletions between the genomes, repeat units that differ between the genomes and inversions (pieces of the genome that are flipped around thus making the genes appear in different arrangements between genomes). I really don’t need to get bogged down into describing all these part of a genome structure here because the methods he used are really not that important. Why? Because the method he used doesn’t really matter to me. I would not deny the he derived this 70% value with the algorithm he used. In fact, does the 70% surprise me? Not really. There are huge tracks (hundreds of millions of base pairs) of the human genome that are repeated sections and pseudogenes (broken, unused genes) and those regions have very low sequence similarity to other animals. With Tomkins way of counting differences I would expect there to be huge numbers of differences. In fact, you and I may only be 80% similar in some regions of our DNA. So what’s the big deal then?
70%, 90%, 98% – what’s in a number?
Well, really percentages don’t matter that much. Who cares if something is 90% similar with a given algorithm or it is 98% similar. What matters much much more is how those numbers compares to other similar pairwise comparisons. In other words, you need to do an apples-to-apples, or oranges-to-oranges if you prefer, analyses to actually be able to compare something meaningfully. So let’s take Tomkins 70% number. That number makes it sound like there is an extreme difference in the DNA sequences of humans and chimps especially when compared to the 96-98% number that has been repeated so much in the literature. But does this mean that humans are so much more different from chimpanzees than scientists have claimed. Honestly, when I began reading this paper the first thing I thought was, this 70% number is completely meaningless unless Tomkins provides me with a context for that number. When I reached the end I found that he provided NO context at all. He continued to talk only of human-chimp comparisons and never once compared his values with the genetic similarities, based on the same methods, of other animals or even among other humans. Without doing so he has done little to nothing to make a case that “the overall extreme discontinuity between the two genomes defies evolutionary timescales.” Why? Because he gave me no idea if 70% really is a lot of difference or not? He is hoping the reader will be astounded by the 70% and never ask, what should be obvious, if 70% really is a lot of difference or does his method of accounting result in all species looking very different?
Tomkins own analysis provides evidence that conflict with his and other creationists’ conclusions about common ancestry!
I asserted above that Tomkins and others are actually undercutting their own beliefs in the genetic uniqueness of man. Where do I get that from? It goes back to the creationist concept of baraminology (See my review of baraminology – Some thoughts on Baraminology). Creationists, such as Tomkins, today are very likely to promote the view of super-speed evolution after the Food. For example, at the AIG Creation Museum, there is a graphic that shows all canine species alive today descending from a pair of dogs on the Ark. Similarly, creationists studying baramins (the original created kinds) have generally concluded that all cats came from a common ancestor. I am writing more about the ostrich and it appears that most creationists think that the ostrich could fly and that the 20 species of extinct Moa’s on New Zealand were once a single Moa that flew to New Zealand and then evolved there. In a nutshell, creationists believe that 100s of thousands of species on this earth are descendants of common ancestors that were found in pairs on Noah’s ark.
Coming back to the 70% value, what does this have to do with baraminology? Tomkins would like his 70% number to sound like such a difference that it would impossible to believe that anything that dissimilar could have a common ancestor. If he believes this is true he could do some very simple analyses to test his hypothesis. His hypothesis could be used to make many predictions but he fails to make any. He could propose tests to see if his 70% number is really a significant value. The biggest test he could propose would be to compare other genomes of other organisms that he believed did descend from a common ancestor. His model would predict that they would show much greater (ie >70%) similarity in their DNA sequences using the same algorithm he used.
When I read Tomkins paper I thought to myself, if chimps and humans are only 70% similar then how similar is a modern human to a Neanderthal or to the Denisovan sequence? For that matter, how similar is one modern human to another. Tomkins provides us with no clue here even though he could have done the analysis. I think this is because he doesn’t want to provide a distraction from his glaring 70% number. I know that those that say that humans and chimps are 96% similar would also find that all modern humans are at least 99% similar to each other and 98% similar to Denisovans and a bit more similar to Neanderthals. Why doesn’t Tomkins apply his same algorithm to other humans sequences and Neanderthals? Since he didn’t it is hard to compare his numbers to other humans and to compare species in other baramins (kinds). What if Tomkins used his algorithm to look at the human and Denisovan genomes and found that they were 90% or even 85% similar. Would he then be able to confidently say that Denisovan’s really were the same as modern humans? What if he compared individual humans and found that that you and I are only 95% similar? Would the 70% value seem as dramatic if modern humans and Neanderthal’s where only 85% similar?
Now here comes the big question: What if Tomkins used the same assumptions for calculating DNA sequence similarity among dog species or cat species and found that they were more different than humans and chimps? What if some of them are only 60% similar? If they are more dissimilar can Tomkins really erect a genetic similarity argument that says that canines all evolved from a common ancestor in 4000 years and are more genetically dissimilar than chimps and humans which he emphatically believes didn’t have a common ancestor? I wonder if Tompkins has ever really thought about the implications of his own data. He has the ability to test his model but makes no mention of what he would expect to find in other species? Why? I get the feeling that Tomkins and others know they are playing with a double standard. They are promoting rapid evolution, including radical genetic change over a short period of time, from common ancestors from one side of their mouth while at the same time proclaiming that genetic dissimilarity precludes common ancestry in other groups from the other side of their mouth.
Of Cats, Dogs and Primates….
I can’t use Tomkins calculations to do my own analyses but based on my own personal experience having worked with a LOT of DNA sequences and done thousands of comparisons of DNA sequences across many groups of organisms I think I have a pretty good idea of what the overall differences between various animals groups would be. So I will make a few predictions right now and then I will collect some data to see if they confirm my suspicions. I will report my findings in my next post on this topic.
I predict that I will be able to find multiple organisms that are demonstrate greater genetic differences than humans and chimpanzees but that
Tomkins and other creationists are likely to believe share a common ancestor. The species that I will target for my analyses will be those in the canine and feline families because creation scientists seem to believe that these two groups each were founded by a single pair of ancestors.
I don’t’ have the resources, nor is there enough data in some cases, for me to compare entire nuclear genomes of these organisms but it is relatively easy for anyone to access entire mitochondrial genomes from thousands of animals and determine their genetic similarity. All animal cells have a separate genome in their mitochondria which is much smaller than their nuclear genome. In humans this genome is 16,569 bases (As,Ts,Cs and Gs). It is similar size in all mammals so it is fairly easy to compare and 10s of thousands of entire genomes have been sequenced. For organisms that aren’t too different from one another differences in their mitochondrial genomes represent a fairly good proxy for the scale of differences I would expect to find in their nuclear genomes (see footnote). Here is something that I do know (based on data from Wikipedia) right now about the mitochondrial DNA of primates:
1) Chimpanzees mitochondrial genomes differ from the average human by 1462 base pairs out of 16,569 total bases in humans. That is an 8.9% difference (or 91.1% similarity). Notice that Tomkins found a 70% similarity for the nuclear genome but the mtDNA genome is easy to compare and so this number is pretty solid and agreed upon by everyone.
2) The average human mitochondrial genome differs from the Denisovan (fossil human) by 385 bp (2.3%)
3) The average human mitochondrial genome differs from the Neanderthal (fossil human) by 202 bp (1.2%)
4) I am not sure how different individual humans are but I’m going to guess that any two individual humans alive today differ by 0 to 75 bp. (ie. less than 0.5%)
I am sure Tomkins would point to the data above and feel confident that it shows that humans and chimps are genetically very different. I would agree, but if he is right that these genetic distance values tell us that humans are not related by ancestry to chimps then if we extend his assumptions to other animals we should expect to see that similar organisms will show much less than an 8.9% difference between them. But here is why I think that Tomkins is in big trouble if he were to actually attempt to do an apples-to-apples comparison of difference species using his methods and assumptions.
Here are some predictions of what I am going to find when I compare mitochondrial genomes within a number of animals groups:
1) Foxes and wolves will be at least as different as chimpanzees and humans and possibly have more differences in the mitochondrial genomes.
2) Domestic dogs will have less variation in their mitochondrial genomes than do humans
3) Domestic cats and/or leopards and cheetahs will be more different than humans and chimpanzees
4) Tigers and lions, which can interbreed, maybe as genetically different as humans and chimpanzees
5) I think that many species of bears may have more differences in their mitochondrial genomes than do chimpanzees and humans. I think the Panda which creationists will say is a bear “kind” will be more different than chimpanzees and humans.
6) Apes, orangutans and chimpanzees which are all placed in the same “kind” by creationist will be more different than one another than chimpanzees will be from humans.
If I am correct in my predictions then it should be apparent that Tomkins and other need to seriously think about how they wish to equate genetic similarity with species boundaries. I’ll start crunching some numbers and come back with a report soon.
Notes: I have done the analysis and and written about it: Of Kinds and Common Ancestors: Comparing Mitochondrial Genomes
DNA sequence similarity: Before someone sends me a comment about DNA similarity not being the only factor in determining genetic relatedness let me assure you that I am aware that two sequences that are very similar can either have the same function or completely different functions. Sometimes just a single change can make a huge difference but in other genes the sequences could be 20% different and the gene would still perform its function in the same way. So gross estimates of genetic similarity are not foolproof ways of assessing the differences between organisms. Since the Tomkins paper is making a case that sequencing differences are too great for there to be common ancestry that is the specific issue I’m interested in exploring right now.
mtDNA: I am fully aware that mtDNA sequences “evolve” at a faster rater than do sequences of nuclear genes.However, compared to the Tomkin’s estimations mtDNA sequences are actually more similar to each other than the nuclear genome. Differences in mtDNA sequences are a rough approximation of how different the whole nuclear genome will be. For example if there is a 9% difference in mtDNA sequence then there is typically going to be about a 2% difference in the coding region of the nuclear genes but if there was 18% difference in the mtDNA there would be bout a 4% difference in the nuclear genome. At much higher rates of differences the differences don’t equate as well but there isn’t enough room here to explain why this is other than to say it has to do with homoplasy for those that know the term.