Understanding creationism, VI:
An insider's guide by a former young-Earth creationist

By David MacMillan.

6. Genetic evidence.

Revised July 4, 2014.

Perhaps one of the clearest and most obvious confirmations of evolution is the convergence between the evolutionary paths of descent determined by fossil evidence and the phylogenetic tree generated by algorithms analyzing genetic information. Because the tree of universal common descent is real, not invented, it leaves the same fingerprint in every part of nature that life touches. Matching trees can be found in global fossil distribution, in analysis of skeletal morphologies, in chromosome length, count, and banding, and in numerous common genetic sequences.

Not every genetic sequence yields a perfect branching tree. Evolutionary theory would not predict perfect branching trees, because random mutations scramble the relationships over time. Even though mutations provide the variation needed for diversification, their accumulation throughout that diversification can eventually obscure the evidence needed to reconstruct those relationships.

Reconstructing phylogenetic relationships is made more difficult because the number of combinations in any given sequence is finite. Every three letters of our DNA codes for only one of twenty different amino acids. Just four or five species with a genetic code as long as our own will have sequences of 60 or 70 base pairs in common – enough to code for simple proteins – simply due to the laws of probability [1]. Such coincidences can also make it more difficult to resolve perfect phylogenetic trees.

It would be a grave mistake, however, to suppose that these difficulties render phylogenetics useless. Just because certain sequences have become corrupted does not mean that a useful tree cannot be constructed. In fact, hopelessly corrupted sequences are the exception, not the norm. When we compare many trees from many different sequences, the accurate phylogeny converges rapidly.

Creationists are fully aware that the match between fossil evidence and genetic evidence is damning. If they match so closely, then common descent must be valid – how else would such agreement be possible? In order to avoid this inevitable conclusion, they seek to invalidate either fossil evidence, genetic evidence, or both, and claim that the apparent convergence is identified only through persistent confirmation bias.

Confirmation bias, of course, takes place when a particular piece of evidence is selected from among many contradictory pieces of evidence because it alone confirms the researcher’s presuppositions. This is almost never a conscious process; someone affected by confirmation bias rarely realizes it. The accusation of bias, then, fits perfectly within the creationist paradigm that mainstream scientists are simply too blinded by their assumptions to see the truth. They don’t even have to accuse scientists of actual dishonesty; it’s all presumed to be part of a nearly innocent, unwitting bias.

Earlier sections have explained how creationists attempt to invalidate the match between the fossil record and the predictions of the evolutionary tree. If they cannot challenge the tree, they will challenge the genetic evidence in whatever way possible.

The most common claim made by creationists unfamiliar with phylogenetics is that the sorting algorithms are somehow “set up using evolutionary assumptions”. This objection reflects a clear lack of understanding. True, the phylogenetic algorithms are set up to produce a branching tree – but the whole purpose of the test is not to establish whether a tree exists but rather to determine whether it is an accurate tree. Creationists imply that there’s some hidden evolutionary guideline built into the algorithm to make it yield the desired result. But that implication is flatly false. The algorithm has no guidance; it has no means of distinguishing between sequences apart from their contents.

Creationists who understand a little more about the subject will instead argue that not all portions of the genome consistently produce the same tree, so researchers are merely picking a tree that just happens to match their expectations. Like many creationist arguments, this simple argument unfortunately makes some intuitive sense. It’s wrong because it doesn’t take statistical probabilities into account.

The more items you have in a given collection, the more ways they can be arranged. Just five items can be arranged in 120 different ways, and ten items can be arranged in a staggering 3.6 million ways. Arranging them in a rooted tree makes the task more complex, so there are many many possible trees; it’s not feasible to simply cherry-pick the one that “happens” to match expectations. In order for any meaningful phylogeny to show up at all, there has to be a legitimate pattern, an actual phylogenetic symbol.

If geneticists were just cherry-picking whatever tree matched their expectations, we would never expect to learn anything from phylogenetic analysis. However, phylogenetic analysis does yield new information. Details are often updated or revised due to the results of genetic studies. The science works.

Moreover, it’s vital to understand that phylogenetic analysis is not limited to one sequence at a time. Phylogenetic analysis is performed on many different gene sequences, allowing researchers to compare results from multiple sources and weed out corrupted data. Corruption is possible, but it never happens the same way more than once, so when multiple sequences generate the same tree over and over researchers can have a high degree of confidence in the result. All of these essentials details about this scientific process are missing from the creationist understanding.

Once they cannot deny that both the fossil record and the genetic evidence are unassailably valid, creationists unveil one more argument: “common design”.

Common design – that morphological and genetic similarities are the result of a designer re-using the same parts – is the perfect creationist argument because it can apply to absolutely anything. No matter how obvious the path of descent is, creationists can simply claim it was intentional. They may also use it in combination with the other objections. For example: “Common design created genetic similarities in creatures with similar environments, similar diets, or similar appearances. These similarities reduce the number of phylogenetic trees to the point that researchers can simply pick whichever one happens to match their evolutionary assumptions.”

The obvious problem is that common design is unfalsifiable. There’s no limit to what it can explain, no level of commonality it cannot be used with. We recognize that an explanation which can fit literally anything is useless; it doesn’t tell us anything. Unfortunately, creationists don’t care whether their explanations are falsifiable. Their presuppositionalist background tells them that it doesn’t matter whether explanations are falsifiable – it’s just necessary to make sure they have the right presupposition at the outset, and everything else flows from that. As long as their denial of mainstream science seems vaguely plausible, they are okay.

So instead of pointing out the unfalsifiability of common design, it’s better to let them use it, but challenge them to take it to its logical conclusion. If their divine common design can really produce the observed levels of genetic similarity, then it should also produce clear and obvious genetic similarities in species that aren’t anywhere close on the evolutionary tree. Not just small sequences in common, but entire gene suites. If God is in the practice of re-using the exact same gene sequences in creatures that happen to show up close together, then we should see the same thing in distant species. Species identified in mainstream science as examples of convergent evolution – the same traits or abilities having evolved separately – should have perfectly matching gene sequences placed there by the creator. For example, bats and birds evolved echolocation separately using different genes, but the “common design” argument would predict the same exact gene sequences.

Any discovery of this nature would be strong evidence for common design. Evolution can explain convergent abilities or small convergent sequences, but not convergent gene suites. Offering creationists this opportunity to demonstrate what they’re claiming puts the onus on them rather than leaving you to try to falsify an unfalsifiable claim.


[1] As mentioned in an earlier section, we have about 3 billion base pairs comprising 1 billion codons, each of which can code for up to around different 20 amino acids. Only 1-2 % of our genome codes for proteins at any given time, so even though the rest of our DNA can participate in the regulation of certain cell functions, it’s pretty malleable. A simple protein may only be a couple of dozen amino acids long, corresponding to about 66 base pairs of length. There are still just under a billion possible 66-base-pair sequences in a 3-billion base-pair genome, meaning that just five different species will have 1036 chances (1 billion to the 4th power) to have two matching 66-base-pair sequences. Since there are 230 possible 66-base-pair nucleotide sequences, those five species will have hundreds of thousands of 66-base-pair sequences in common.