In a previous thread I discussed what science meant by random mutations and how this term is defined by experimental results and statistics. A lot of this work was done in the 1940’s and 50’s, and since then many new facts have been discovered in the field of genetics, such as the discovery of DNA. This gave rise to the field of molecular biology and allowed us to understand biology at the molecular scale.
So what exactly causes mutations? We once again have to reiterate the the limited scope of science. What science can do is put forward hypotheses and see if the evidence is consistent with that hypothesis. What science can not do is make ontological statements about absolute truth. With that in mind, the next few posts will discuss the evidence that links mutagenesis (the production of mutations) with the biochemistry of the cell and why scientists aren’t simply assuming that mutations are caused by biochemistry.
First off, a ton of credit goes to @glipsnort for his wonderful essay over at BioLogos.
I could be wrong, but I think @glipsnort also cited a Paabo paper that demonstrated the same relationships which first got me interested in looking into this data a bit more. Anyway, big props to Steve Schaffner. Like the cited essay, I will be focusing on substitution mutations which involves switching out one base for another. We could also discuss insertions and deletions, but that is a far more technical discussion than substitutions.
The first concept we need to look at is the DNA bases themselves. There are two types of DNA bases: the purines and the pyrimidines. To put it simply, two of the bases have one ring and two of the bases have two rings (three if you count uracil which is found in RNA).
Bases that are more similar to each other have a higher chance of being switched out for each other. The proteins that copy DNA are going to be more likely to mistake an adenine for a guanine, or a cytosine for a thymine. When there is a switch between similar bases it is called a transition mutation, and a switch between dissimilar bases is called a transversion mutation.
There is also another biochemical process that biases mutations which are CpG mutations. CpG is a short way of saying cytosine—phosphate—guanine. For example, there is one CpG in this sequence: AATGGACGTAATT. Methylases like to put a methyl group on the C of a CpG, and this makes the C susceptible to further chemical modification into a T.
These are the two main biochemical processes that bias certain substitution mutations over others. Contrary to popular belief, random mutation does not mean that all mutations are equally likely. Instead, some mutations are more likely than others.
The next step is to see how this plays out in living populations.
One way of detecting mutations is to sequence the genomes of parents and their offspring. In this example, 78 family trios were surveyed for mutations, and these are the results:
First off, we can see that transitions at non-CpG sites is about twice that of transversions at non-CpG sites for rate per base per generation (the column on the right). Since CpG’s are very susceptible to mutation they are counted separately. CpG transition mutations are about 20 times more common (on a per base basis) than transitions at non-CpG sites. As you will notice, there are two figures: number and rate. Even though CpG transitions have the highest rate there are relatively few CpG’s in the genome compared to single A’s or T’s, for example. This is why we need to look at both the number and the rate.
As we would expect from our knowledge of biochemistry, transition mutations should outnumber transversion mutations and CpG transition mutations happen at the highest rate. This is observed in living populations.
Next, we will use these observations to test hypotheses about common descent and how evolution explains the differences between species.
The hypothesis we are testing is straightforward: is biochemistry responsible for the differences between individuals within a species and between species? To test this hypothesis we look for the same pattern of transition, transversion, and CpG mutations.
“Genomewide average frequencies for various nucleotide differences between chimpanzees and humans ( A ) and among humans ( B ). Ti = transitions.” Ebersberger et al. (2002)
We can see the same pattern in the chart above. (B) is the comparison between human genomes, and we see that the CpG mutations dominate. (A) is the comparison between chimp and human DNA, and we see the same pattern of specific types of substitution mutations as those seen in the comparison between humans.
@glipsnort (aka Steve Schaffner) has a very similar chart that excludes the CpG mutations so that we can see the relationships between the types of mutations with lower rates:
Again, transitions outnumber transversions in both the human-human comparison and the human-chimp comparison, and the same trends are seen in both.
This also extends to comparisons of more distantly related species. I will leave out the graphs here to keep the clutter down, but you can check them out here.
The evidence supports our hypothesis. The evidence is consistent with biochemistry and common descent producing the differences between genomes. Does this mean that God is not guiding these mutations? NO!! Science has no way of proving or disproving this claim. What we can say is that the evidence is consistent with our hypothesis. Period.
It is this type of evidence that has led scientists to conclude that mutation is the cause for differences between species. This isn’t assumed. This is a supported conclusion. Evolution explains why we see these differences as much as any other scientific theory explains how nature works.
Excellent thread @T_aquaticus. Thank you for pulling this together. This is exactly the sort of evidence that convinced me that evolution was not merely an “argument” (see @mercer?), but offered effective and quantitative explanation of intricate details of biology. This was a watershed realization for me, because none of the ID or YEC argument learned ever taught me these things. It was not as if they had a solid rebuttal, they just did not know or did not engage with this sort of strong evidence.
More than a decade ago I asked myself why all of these people could be creationists. I thought that maybe, just maybe, there was something to it. After reading creationist materials and looking at what they presented I came to the same realization you did. Creationism is extremely superficial and it doesn’t dig down into the biology. Many of these pieces of evidence have been understood for a long time, and yet creationism has failed to address them. I don’t see any path that would allow creationism to explain this type of data. Of course, I would be absolutely giddy to be proven wrong.
It is rather frustrating to see creationist claims like “evolution doesn’t explain the differences” or “scientists just assume they are mutations”. I am hoping that threads like these will at least help some creationist understand why scientists have arrived at certain conclusion.
That is exactly the right attitude for every scientist.
This is one of those pieces of evidence, though, that is just so strong that I wonder, at times, if they just don’t want to give it publicity. They do not have a good argument against it, and it is not widely known outside of science, so it is better just be quiet about it and hope no one notices.
For me, this backfired big time. It took just a 15 minute conversation to overturn over a decade of indoctrination. I was walking with a couple biology professors as an undergrad, asked them why they thought common descent was true. They went straight to the evidence like this, explained it briefly. I checked it out myself. There was not a single response to it by ID or YEC or OEC creationists any where, and there still is not. Of course,I knew why immediately. It is a smoking gun for common descent, and it is not theory laden, making no assumptions.
That is when I realized that, at the very least, genomes demonstrated that humans appeared to share common ancestors with the great apes. God could have made us so it did not look this way, but He did not. Why not? That was the question I took with me into graduate school. That wasn’t enough to bring me over, but that was the beginning of the end of my scientific anti-evolutionism. It was like pulling back the curtain to see the Wizard of Oz.
Statistical Quibble: I have never seen a paper report a p-value with an exponent 10^{-300}. In Frequentist stats we only care if the p-value is smaller than the stated alpha-level, after controlling for multiple comparisons.
This is why it is a bit frustrating to hear rhetoric like “you have to accept evolution as a biologist in order to keep your job, and that is why so many scientists accept the theory”. It simply isn’t true. Once you understand the basics of biology and genetics the evidence smacks you in the face like a 15 ton truck. A lot of times it comes down to the old saw of leading a horse to water. You simply can’t force people to drink. All you can do is present the evidence and hope that their curiosity causes them to consider the implications.
@T_aquaticus, I found the information presented in your post and in @glipsnort 's article on Biologos to be very interesting and strongly indicative of common ancestry. So thank you for putting it together.
Of course, I’ve managed to come across someone who disagrees in a conversation on Facebook… he thinks the data in the Kong et. al. article that you link here actually contradicts the data laid out in the post on Biologos. Here’s what he said:
Now, my suspicion is that this objection isn’t hard to overturn if one has the relevant knowledge, but alas, I do not have it. (I can only say that I don’t know enough to determine whether the two datasets are statistically compatible, though the fact that the relevant ratios are the right order of magnitude suggests to me that they might be.) Anyone here wants to take a stab at a response?
For a start, it should be noted that Kong et al. (2013) counted less than 5000 mutations in total, which isn’t a big enough sample size to get really accurate estimates of the relative proportions of the different types of mutations. It gives a “low-resolution” look at relative mutation rates, of the right orders of magnitude, but shouldn’t be relied on for more than that. More mutations have to be counted to get more precision.
Fortunately, this work has already been done. In my blog post on the subject (linked below), I gathered data from 3 studies: Francioli et al. (2015); Wong et al. (2016); and Halldorsson et al. (2019), who between them counted about 220,000 de novo mutation mutations, compared to Kong et al.'s mere 4933. When comparing the spectra from these counts with those from human diversity and human-chimp differences, I found a near-perfect match:
@structureoftruth what does the Facebook objector say in response? I found it important and interesting that he seems to understand the argument, so this should put his objection to rest. Does it?
What that person missed is the graphs are based on rate, not count. To calculate the rate you take the count and divide by the number of places where that mutation could occur. The number of CG’s in the human genome is less than the number of single A, T, C, or G. Therefore, there are fewer places where CpG mutations can occur. Also, there are more A’s and T’s than C’s and G’s in the human genome. Therefore, the rate of mutation takes out the bias introduced by the different number of each base (or combination of bases in the case of CG) in the genome.
But (non-CpG) transitions and transversions can happen to any base (A, T, C, or G) in the genome, so how would the relative numbers of any of the bases affect the relative numbers of transitions vs transversions?
