Vitellogenin and Common Ancestry

In my role at BioLogos, I frequently explore the diverse and extensive evidence for common ancestry – the evolutionary idea that species we see in the present day share common ancestors in the past. One such line of evidence is the presence of mutated genes that no longer perform their original function. These sorts of sequences, called pseudogenes, are widespread. Moreover, for many of these defective genes, we share identical mutations with other species – most often with chimpanzees, then with gorillas, then with orang-utans, and so on. The pattern of these shared mutations forms what is known as a nested hierarchy. Shared mutations in pseudogenes that form nested hierarchies are an excellent way to determine how species are related to one another. For example, prior to genome sequencing that provided these data, there were a few scientists that maintained that gorillas were a closer relative to humans than chimpanzees. They based their argument on anatomical evidence, though their views were very much a minority view. Genome sequencing helped resolve this question, in part by examining shared mutations in pseudogenes.

The nose knew

An example that illustrates this well is olfactory receptor pseudogenes. Olfactory receptors are proteins present on nasal surfaces (the olfactory epithelium) that bind on chemicals in the air we breathe, change their shape in response to that binding, and then use their changed shape to transmit signals to our nervous system that we perceive as smell. Humans, chimpanzees, and gorillas have lost the function of many olfactory receptor genes, and we share many identical mutations in common with these other apes. If humans are more closely related to chimpanzees, this means that we share a longer common ancestral lineage with them than we do with gorillas. If we are more closely related to gorillas, the converse is true. These two hypotheses predict one of two patterns for olfactory receptor pseudogenes. If we are closer relatives to chimpanzees, we should observe some mutations in olfactory receptor pseudogenes that we share with chimpanzees but not with gorillas. Conversely, if we are more closely related to gorillas, we should see some mutations that are shared between humans and gorillas but not with chimpanzees:

Figure 1: When we examine genome sequencing data, we see many identical mutations shared by all three species, but more importantly some mutations shared between humans and chimpanzees that gorillas do not have. This evidence (which also happens to line up with all other genome sequence evidence) shows us that chimpanzees are our closest relatives.

You’ve come a long way, baby

While it’s not too surprising to see olfactory receptor pseudogenes in mammals, pseudogene sequences can persist for a very long time in a lineage – long enough that their presence can become incongruous with the lifestyle of the organism that harbors them. One example that I have discussed several times in the past is the curious case of vitellogenin pseudogenes in placental mammals. Vitellogenins are large proteins used by egg-laying organisms to provide a store of nutrition to their embryos in egg yolk. Since vitellogenins are so large, they are a good source of amino acids when digested (proteins are made of amino acids linked together). Many of the amino acids in vitellogenins have sugars attached to them as well, so they also serve as a source of carbohydrates. The three-dimensional shape of vitellogenin proteins also acts as a carrier for lipids. As such, vitellogenins can be synthesized in the mother and transferred to the yolk as a ready-made supply of amino acids, sugars, and lipids for the developing embryo.

Placental mammals, on the other hand, use a different strategy for nourishing their embryos during development: the placenta. This connection between the mother and embryo allows for nutrient transfer right up until birth. As such, there is no need for vitellogenins, or storing up a supply in the egg yolk for the embryo to use. Evolutionary biology predicts that placental mammals descend from egg-laying ancestors, however – and one good line of evidence in support of that hypothesis (among many) is that placental mammals, humans included, have the remains of vitellogenin gene sequences in their genomes.

Looking for Answers

Not surprisingly, this evidence is problematic for those holding an anti-evolutionary perspective. Back in 2012, I surveyed the websites of anti-evolutionary groups to examine how they handled this evidence – but came up empty-handed. Though young-earth groups (such as the Institute for Creation Research and Answers in Genesis), old-earth groups (such as Reasons to Believe), and supporters of Intelligent Design (such as the Discovery Institute) all had articles discussing pseudogene evidence in general, as well as the particular case of the GULO pseudogene that makes humans and other apes dependent on dietary vitamin C, none specifically dealt with vitellogenin pseudogenes in particular:

Interestingly, this situation has now changed. Though the Discovery Institute and Reasons to Believe remain silent on the topic, a paper has now been published by young-earth creationist (YEC) researcher Jeffery Tomkins that claims to debunk the vitellogenin pseudogene story. The paper is published in Answers Research Journal (a journal owned and operated by Answers in Genesis), though Tomkins works for ICR. In a blog post entitled “Evolutionists Lay an Egg: Vitellogenin Pseudogene Debunked.” Tomkins introduces his work on the ICR website as follows (where he abbreviates vitellogenin as “vtg”):

One study claimed to have found genetic evidence of an ancient vtg gene in the human genome. Because the actual data for this discovery were questionable, the evolutionary community in general did not actively popularize the alleged finding. However, BioLogos, a religious group of evolutionary scientists and liberal Christian theologians, has been promoting the so-called egg-laying pseudogene discovery as evidence of evolution. Because this type of propaganda is targeted to the Christian community, the claim should be more thoroughly investigated.

Tomkins goes on to claim that he has overturned the evidence that the human genome has a vitellogenin pseudogene, claiming rather that it is part of a functional gene. As such, he claims the evidence shows that the “so-called egg-laying pseudogene” is not evidence for evolution, and that BioLogos is in error in presenting it as such. Since I am the primary BioLogos author who uses this example, and since I am committed to representing scientific evidence as accurately as possible for the BioLogos community, I was naturally interested in Tomkins’ claims. Do I have vitellogenin on my face, as it were?

In a nutshell, I do not think the case Tomkins makes stands up to scrutiny – but explaining why will require some effort. As we undertake a careful analysis of his work, we will also have opportunity to delve more fully into the science related to how placental reproduction came to be within our lineage – an area of biology that is fascinating in its own right.

In the following section, we’ll start by examining the vitellogenin sequences found in the human genome, and begin to discuss the lines of evidence that support their evolutionary history.

In C.S. Lewis’s book Prince Caspian, the Pevensie children find themselves, upon their unexpected return to Narnia, in the ruins of their former palace, Cair Paravel. Unaware that they are several hundred years beyond “their time” as kings and queens as recounted in The Lion, the Witch, and the Wardrobe, they at first do not recognize the ruins. Over time, however, they begin to notice similarities. Though Cair Paravel was not on an island, nor was an apple orchard planted up to the walls, there were too many similarities to ignore. The great hall, with its dais, was the correct size and shape, though now roofless. The well was in the correct location, and also the correct shape. Near the well, Susan had found a gold chess piece, exactly alike to the ones they remembered using. Though at first difficult to imagine, more and more pieces fit into place, strengthening their conviction that they are, in fact, in the ruins of their once glorious home. As they begin to “see” it as it was, Lucy hits upon an idea that will settle the matter readily:

‘There’s one thing,’ said Lucy. ‘If this is Cair Paravel there ought to be a door at the end of this dais. In fact we ought to be sitting with our backs against it at this moment. You know – the door that led down to the treasure chamber.’

‘I suppose there isn’t a door,’ said Peter, getting up. The wall behind them was a mass of ivy.

‘We can soon find out,’ said Edmund…

–C.S.Lewis, Prince Caspian

As those who have read the story know, the children soon discover beyond doubt that they are indeed in the ruins of Cair Paravel: they find the treasure chamber, and within it the gifts that Aslan had given them hundreds of years before.

The reason that Cair Paravel remained recognizable, even after hundreds of years, was that several of its features were resistant to the effects of long spans of time. Though the original peninsula could become an island, an orchard overgrow, and a roof fall in, other features would not shift: the shape, size, and relative placement of the walls and rooms persisted.

Like buildings, the DNA sequences in an organism’s genome can retain features for very long spans of time even after their original function is lost. These features can be used to make predictions about what we should find, and where. One such feature is the organization of genes along a chromosome. For example, consider two closely related species. Since these species recently shared a common ancestral population, and thus the same genome, we would expect the majority of genes in these two species to remain in the same locations relative to each other (Figure 2):

Figure 2. Closely related species have not only highly similar genes and inter-gene sequences, but have their genes conserved in the same spatial arrangement: in other words, they have large blocks of genes with shared synteny.

This conserved gene order along chromosomes is known as shared synteny: “syn” here means “together” and “teny” means “held”, so “synteny” means “held together” – genes held together in the same pattern in related organisms. Genes in closely related organisms can be present in blocks of synteny that are thousands of genes long. Over time, gene order can be rearranged through chromosome breakage and rejoining – slowly erasing shared synteny in separate lineages. Despite these processes, however, shared synteny can persist for a very long time in distantly related species. Like the ruins of Cair Paravel, it takes a long time to remove the overall pattern to the point where it is not recognizable.

Going on an egg hunt

The knowledge that shared synteny can persist in separate lineages for a very long time was useful when looking for vitellogenin gene fragments in the human genome. As we saw in Part 1 of this series, placental mammals do not require vitellogenin genes to supply their embryos with yolk. Converging lines of evidence, however, indicate that placental mammals are the descendants of egg-laying ancestors. For example, placental mammals and birds (a group of modern-day, egg-laying organisms) are thought to have shared a common ancestral population about 310 million years ago. If this is the case, then it is possible that vitellogenin gene sequences, however fragmentary, might remain in the human genome. The scientists interested in this question used synteny to find the regions of the human genome worth examining.

Modern birds (such as chickens) have three vitellogenin genes: VIT1, VIT2, and VIT3. The latter ones sit side by side in the chicken genome, with VIT1 in a different location. The three VIT genes sit next to other genes in the chicken genome: VIT1 sits next to a gene called “ELTD1”, and VIT2 and VIT3 sit between genes named “SSX2IP” and “CTBS”. These genes are not involved in making egg yolk – they just happen to be the closest neighbors of the VIT genes (Figure 3):

Figure 3. The genomic context for the three VIT genes in chicken.

With these data in hand, the researchers then searched the human genome for the genes near the chicken VIT genes. These three genes (ELTD1, SSX2IP, and CTBS) are also found as functional genes in humans – and as expected, these genes have the same spatial arrangement in the human genome as they do in chickens (Figure 4):

Figure 4. The genes flanking the VIT genes in chicken are present in the human genome in the same spatial pattern.

The researchers then did a careful sequence comparison between these regions in the human and chicken genomes. The gene sequences for ELTD1, SSX2IP, and CTBS were already known to be highly similar, but they found that other sequences in this region matched as well. We can represent sequence matches between the two genomes as a black bar between them to show what they found (Figure 5):

Figure 5. Shared synteny between humans and chickens spanning the regions with functional chicken VIT genes. Black bars between the two chromosomes indicate sequence matches. The match between humans and chickens is statistically significant at the site of VIT1, but the fragments at VIT2 and VIT3 are too short to support statistical significance. This figure is based on data from Brawand et al., 2008.

What is important to note is that not only did the researchers find evidence of fragmentary remains of all three VIT genes in the human genome (though the human VIT1 sequence was the best preserved of the three) they also found sequence matches that span both regions on either side of the genes in question. This evidence increases our confidence that we are indeed looking at regions with shared synteny: in other words, a region in two present-day species that was once a region in the genome of their common ancestral population. The human VIT sequences, as expected, are far too fragmentary to act as functional VIT genes.

In the following section, we’ll begin to explore how Jeffery Tomkins examines this evidence in his attempt to refute it.

When I was a young person growing up in northern British Columbia, one area of my hometown had a significant number of houses that were nearly identical to one another. These houses were prefabricated and erected at the same time – spanning about two blocks. As it happened, I knew two families that lived in this area, and so spent time in both of their houses. Once you knew your way around one house, you knew your way around the other. Despite small differences like color and décor, the two houses were the same: the same layout, with the same rooms.

Over time, the houses in this area became more distinct from each other – some have had additions put on, others have had rooms remodelled (and I’m certain this process has continued to this day, though I haven’t been back to my hometown for many years). Despite the changes, these houses remain recognizable as coming from the same pattern – and examining a number of houses for their common features shows what that original pattern was.

In the previous section, we examined the evidence for shared synteny between the human and chicken genomes for the regions with functional vitellogenin (VIT) genes in chickens. (Recall that shared synteny is merely the technical term for two or more genes in the same spatial pattern – i.e. order along a chromosome –  in two or more organisms). What we saw in Figure 5 is that there is good evidence that the regions containing VIT pseudogene fragments in the human genome once shared common ancestral sequences with the VIT regions in chickens: the regions have the same sequences in the same overall pattern, even though changes have occurred since these sequences went their separate ways:

In one sense, this is like looking at two houses in that neighbourhood in my hometown. Seeing the same features in the same pattern indicates that these sequences were inherited from a common ancestral population – just like seeing the same structures in those houses indicated they were part of the original construction.

But, you might ask – what about the other houses?

Evolutionarily speaking, the observed shared synteny for the VIT regions in humans and chickens makes a prediction about what we should find in other mammals. Since the last common ancestral population of humans and chickens lived prior to the evolution of all mammals, we would expect to (at least potentially) find these regions in any other mammal we care to sequence – with the understanding that these sequences might be missing if they have been lost in a particular lineage. To return to the house analogy, we would expect shared features unless renovations in any particular house obscured them.

The researchers thus looked for VIT genes in a diverse number of mammals, and, not surprisingly, found them in the same arrangement as seen in chickens and humans. One example comes from a marsupial mammal – the opossum. Just like for humans, the opossum VIT genes are riddled with mutations that prevent them from being translated into proteins. Despite those mutations, however, enough VIT gene remnants and their non-gene flanking sequences remain in opossums to easily identify them – nested between the same functional genes we see in humans and chickens. In fact, in opossums, more of the VIT2 and VIT3 sequences remain than in the human genome, and more of the DNA flanking VIT1 remains the same:

Figure 6. Shared synteny between opossums and chickens spanning the regions with functional chicken VIT genes. Black bars between the two chromosomes indicate sequence matches. This figure is based on data from Brawand et al., 2008.

These findings, then, match what common ancestry predicts if indeed humans, chickens, and opossums share a common ancestral population deep in the past. Opossums, since they do not lay eggs, do not require VIT genes any more than placental mammals like humans do. Nonetheless, they too have remnants of these genes in the exact places in their genomes that common ancestry would predict. Moreover, the researchers found that several other placental and marsupial mammals also have VIT pseudogenes. As you might expect, however, egg-laying mammals (such as the platypus) retain a functional VIT gene that they use to perform bulk yolk transfer to their embryos.

In summary, what we see is a broad pattern of evidence that supports the hypothesis that placental and marsupial mammals share common ancestral populations with egg-laying mammals, and more distantly, other egg-laying vertebrates such as birds. This hypothesis was originally proposed based on shared anatomy and physiology, but has now been tested down to the molecular level – and has passed with flying colors.

Reading Tomkins

With this context in mind, we are now ready to evaluate how a recent paper, written by a young-earth creationist and published by Answers in Genesis, attempts to rebut this evidence. As we saw in the first post in this series, Tomkins claims to have “debunked” this example – and shown it not to be evidence for evolution.

How Tomkins claims to achieve this in his paper is key: he focuses only on one fragment of one of the VIT pseudogenes in the human genome. This fragment is the largest continuous fragment of the human VIT1 sequence at about 150 nucleotides long. Tomkins describes it in a blog post as follows:

The main piece of evidence for the vtg pseudogene is the presence of a 150-base human DNA sequence that shares a low level of similarity (62%) to a tiny portion of the chicken vitellogenin (vtg1) gene. However, the chicken vtg1 gene is actually quite large at 42,637 bases long, so a 150-base fragment of 62% similarity represents less than 0.4% of the original gene if the evolutionary story were true!

In the paper itself, Tomkins claims that this fragment is the extent of the VIT1 pseudogene in humans:

The sequence identified by Brawand, Wahli, and Kaessmann (2008) as being a vtg pseudogene is only 150 bases long…

Note well: this is the only sequence that Tomkins will address in his paper (!): not the other VIT1 sequence remnants surrounding this fragment; not the shared non-gene sequences that flank VIT1 in humans and chickens; nor even a mention of the VIT2 or VIT3 regions with their pseudogene fragments, nor the flanking DNA also found there. Similarly, the finding that these regions are shared with a wide array of other mammals is not mentioned. Tomkins has neatly bypassed the bulk of the evidence with this approach by removing the one fragment he discusses from its context, and ignoring the VIT2 / VIT3 region altogether.

The true “main evidence” for the remains of VIT genes in the human genome is as we have discussed: the overall match of sequences between placental / marsupial mammals and egg-laying organisms over large spans of DNA, including flanking regions. This is the evidence that needs to be addressed – and Tomkins does not even mention it, let alone address it. It is also highly unlikely that his audience – since Tomkins is writing not for biologists but rather for laypeople who follow young-earth creationism – will be able to see this problem in Tomkins’ approach. Moreover, since Tomkins tells them that this fragment is the extent of the VIT1 pseudogene, they would have to read the original paper by Brawand and colleagues to notice this is incorrect.

In the next section, we’ll see that even Tomkins’ treatment of this small VIT1 fragment has several problems.

In the last post in this series, we saw how Tomkins – in his attempt to refute the striking evidence for common ancestry that human vitellogenin (VIT) pseudogenes provide – chose to ignore the entirety of the evidence except for one fragment, 150  base pairs long, of the human VIT1 pseudogene. From this starting point, he then attempts to convince his readership that this fragment is not a VIT pseudogene fragment at all, but rather part of a gene with a function unrelated to egg yolk formation.

Even as we turn to evaluating that argument, it’s important to remember that doing so is secondary to what we have already addressed: Tomkins has chosen to ignore almost all of the evidence for human VIT pseudogenes, and all of the evidence for VIT pseudogenes in other placental and marsupial mammals. Even if Tomkins’ case for the one fragment he chooses to address was airtight, it would not even begin to be a satisfactory rebuttal of the evidence for a scientifically informed audience.

Tomkins on shared synteny

As we have seen in prior posts, the shared synteny between the human and chicken genomes for the VIT1 region is an important piece of evidence for their common ancestry.  Though Tomkins is aware of the importance of this line of evidence, he addresses it dismissively:

One of the supporting arguments for the vtg pseudogene fragment being authentic is that it shares gene neighborhood synteny with chicken. When this was investigated, it was found that gene synteny surrounding the chicken vtg1 gene (~360,000 bases) compared to the region surrounding the alleged vit1 fragment in human, was completely different except for the presence of the LTD1 [sic, what Brawand et. al call ELTD1] gene which was about three times the distance (~100,000 bases) from the alleged vit fragment as its homolog is in chicken (~38,000 bases). [Emphasis mine]

Tomkins is recognizing that the ELTD1 gene is the same in humans and chickens, but he claims that the rest of the surrounding area is “completely different” and that it is further away from VIT1 in humans than in chickens. Of course, admitting that the ELTD1 gene sits beside the functional VIT1 gene in chickens and the VIT1 pseudogene fragment in humans is a concession to shared synteny – but notice how Tomkins treats this evidence. His primary claim is that aside from the ELTD1 gene sequences, these regions are “completely different” in humans and chickens. (Additionally, he points out that the number of base pairs between the human and chicken sequences and ELTD1 differs – though I am unclear why he would think this supports his case, unless he is not willing to consider that insertion and/or deletion mutations in this region may have altered the distances over time.) His primary claim – that these regions are “completely different” is incorrect, as we can see from the data in the original paper, and diagrammed here:

Figure 7: ABOVE: Shared synteny between humans and chickens spanning the VIT1 region. Black bars between the two chromosomes indicate sequence matches. There are several matches between the functional VIT1 in chicken and the human genome, though the human VIT1 sequences are fragmentary. This figure is based on data from Brawand et al., 2008.

Not only are there several sequence matches in the region between the VIT1 fragment and the ELTD1 gene, some of those sequences are VIT1 pseudogene fragments that Tomkins is ignoring. Additionally, there are matches that extend beyond the ELTD1 gene and the VIT1 fragments, which establish that these sequences share synteny.

I am not exactly sure why Tomkins would claim that these sequences, which do indeed have matches, are completely different. It may be that they are “completely different” under the particular (and highly idiosyncratic) methods he has used to do genome comparisons in the past, for which he has been criticized, even by other young-earth creationists and supporters of Intelligent Design (see here and here, for example). In any case, Tomkins’ claim in this instance is simply wrong, and would greatly mislead a non-specialist audience.

Pseudogene or functional gene?

Having rebutted the synteny evidence to his satisfaction, the rest of the case that Tomkins attempts to make is straightforward to understand: He argues that the 150 base pair fragment of VIT1 is “not a real pseudogene” but rather regulatory DNA for an unrelated gene – DNA that functions to direct where and when another gene is transcribed. This gene falls into a class of genes known as “long non-coding RNAs” – genes that are transcribed into RNA but do not code for proteins. These genes are thought to arise quite easily over the course of evolution, and many are thought to have little or no function (for a good but technical review, see here). Specifically, the VIT1 fragment Tomkins acknowledges sits in an intron of this gene (part of the gene that is spliced out of the RNA copy after it is transcribed from DNA). Tomkins summarizes the evidence – such as it is – that suggests this sequence might have a regulatory function. If this fragment is a functional part of an unrelated gene, he argues, it makes sense to understand it as the result of special creation, rather than the evolutionary remnant of a once-functional VIT1 gene in the lineage leading to humans.

The major problem with this argument is that it subscribes to a false dichotomy: that this sequence is either a VIT1 pseudogene fragment or a functional part of another gene. From an evolutionary perspective, there is no issue with it being both. Part of evolutionary theory is the expectation that occasionally some sequences, after losing their original function, may come under natural selection to be repurposed to another function. The technical term for this process is exaptation, and many examples of it are known. Certainly a long, non-coding RNA gene could arise at this location in the human genome and this sequence could be exapted as a regulatory sequence – but there is no hint of admitting this possibility in Tomkins’ work. Once again we are reminded that he is not writing for a scientifically informed audience, but rather for a lay audience that will not know about the possibility of exaptation and expect him to provide evidence that it is not a factor in this situation. Rather, it seems enough to Tomkins to suggest that the sequence is functional – and that this alone will be enough for him to convince his readership that this fragment is “not a real pseudogene.”

Cair Paravel, prefab houses, and exaptation

Perhaps returning to our prior analogies of synteny will be helpful here. Recall that when the Pevensie children (of Chronicles of Narnia fame) were coming to the realization that they were in the ruins of Cair Paravel, they had to clear away a mass of ivy from a wall before they could access the door to the treasure chamber. This wall had previously held up the roof of the great hall above the dais, but now it was a support for the ivy, and undoubtedly a host of animal life within it. The “function” of the wall, then, had changed over time. None of the children, of course, pointed to this “new function” as incompatible with their hypothesis that the wall had formerly been used for a different purpose. The observation that the wall was now a home for plants, insects and perhaps even birds was not a problem, since the children understood that these new properties could have been accumulated over time as the original function was diminished. Moreover, they correctly understood that the weight of the evidence around them strongly suggested that this transformation indeed had taken place.

Our second analogy was the neighborhood of prefabricated houses in my hometown – originally constructed as identical, and gradually becoming more distinct over time as each house has its independent history shaped by renovations and alterations. Suppose in this neighborhood one homeowner decided to relocate the bathroom into an addition, and repurpose the original bathroom into a laundry room. Careful examination of the laundry room might show how the plumbing was adapted from its original function to serve in this new capacity. Imagine, for example, that the drain for the toilet remained in its original position, but was capped and recessed under the floor. Though the room might have been converted to a new function, it still would retain some features that indicated its prior history – and examining other houses in the neighborhood would only confirm our suspicions. Merely discussing how the washing machine was hooked up properly (perhaps where the sink had once been) would do nothing to convince us that the toilet drain under the tiles was anything other than the clear sign the room had once been a bathroom.

Put more simply, evidence of function does not erase the evidence for prior history.

Even though the evidence that this sequence is functional in humans is rather thin, the main issue is that even if its function were convincingly demonstrated in the future, it would not remove the evidence that this sequence was once part of a functional VIT gene – evidence that Tomkins has either not addressed, or denied outright.

In the next section, we’ll leave Tomkins behind and delve into the biology of how a lineage might shift from laying eggs to placental reproduction – a shift that has also occurred in lineages other than the one leading to placental mammals.

In the first four sections of this article, we have seen that there is strong genetic evidence to support the hypothesis that placental mammals descend from egg-laying ancestors. One such line of evidence is the remains of three vitellogenin genes in the genomes of placental mammals exactly where common ancestry predicts them to be. Vitellogenins, as we have seen, are necessary for forming egg yolk. In placental mammals, yolk is not made, and embryos are connected to the mother to receive direct nourishment throughout development.

Often when I present on these lines of evidence, questions arise: how is it that an organism could suddenly stop laying eggs and switch to live birth? How could a placenta suddenly appear? Many of these questions are rooted in the misconception that evolution is a sudden process, perhaps requiring large mutations that dramatically alter one organism. Rather, evolution is a gradual process where average characteristics of a population shift over time. As such, evolution predicts that the transition to live birth and a placental connection was a gradual one. Even with these understandings in place, however, the gap from one form of reproduction to the other seems too far for evolution to bridge.

As Tompkins puts it:

The grand evolutionary story claims that egg-laying creatures share a common ancestry with placental mammals. Non-mammalian vertebrates, such as birds and reptiles, lay eggs with nutritional reserves in the egg yolk to nourish the growing embryo inside. In contrast, the embryos of placental mammals are nourished through the placenta, a specialized organ attached to the uterine wall of the mother. Placental mammals are born alive and do not hatch from eggs.

The supposed transition from an egg-laying reproductive system to a placenta-based system is notoriously difficult for evolutionists to explain.

So, is this transition “notoriously difficult” to explain, as Tompkins claims? No, actually, it’s not. Let’s take a look.

Placental reptiles?

One thing to keep in mind in this discussion is that mammals, as a group, are nested within reptiles. In other words, mammals are modified reptiles. There are numerous transitional fossils linking reptiles to mammals, and the now out-dated names for these groups were once “mammal-like reptiles” and “reptile-like mammals” with no clear dividing line between them, illustrating the gradual shift from reptile to mammal.

So, present-day reptiles and present-day mammals are close relatives. Hold that thought.

Those who have read my Evolution Basics series will also be familiar with the idea of convergent evolution – the observation that evolution often “repeats itself” in lineages. We see fish and dolphins arrive at very similar body shapes, for example – shapes that were not inherited from their last common ancestral population.

One way to explore the possibility of an egg-laying to live-birth to live-birth-with-placental-development transition in the placental mammal lineage is to see if other lineages have also – convergently – arrived at something like placental development from an egg-laying starting point. The place to look for such transitions would be among mammalian relatives, such as reptiles.

When we do this, we find numerous examples of reptiles that have shifted from egg laying to live birth. We also find examples of live-bearing reptiles that have placentas. Moreover, there are even examples of reptiles (skinks, as it happens) of the same species that can either lay eggs, or bear live young, depending on the environmental conditions they find themselves in. In challenging conditions, they retain their embryos for live birth. In easier conditions, they lay eggs. All that is required for live birth in these “dual mode” species is an eggshell that progressively becomes thinner and thinner as development proceeds, such that it is merely a thin membrane at the time of birth.

Ironically, the young-earth creationist organization Answers in Genesis has commented on these observations, and concluded that the ability to switch between egg laying and bearing live young in reptiles (and the various types of placentas seen in reptiles) is consistent with their view of “created kinds”:

Extant live-bearing lizards and other reptiles display an array of placental morphologies. Furthermore, at least two species of skinks contain both egg-laying and live-bearing members. While touted as evidence for evolution, these species are actually good illustrations not of evolution but of genetic variation within created kinds. Those dual-mode skink species exist in different climates, each variety being well-adapted to the challenges of its environment. And there is no evidence that those skink populations are changing their reproductive habits at all or evolving into new kinds of creatures.

While we cannot know for certain how many created kinds of lizards God made in the beginning, we could postulate that the genetic ability for egg-laying and live-birth co-existed at least in skinks of the dual-mode types …  the diversity of placental morphology among reptiles is consistent with the biblical concept that the created kinds of creatures diversified to fill the ecological niches of the world. Natural selection (and other mechanisms) would have allowed the best-adapted populations to survive in each habitat.

So, from a young-earth perspective, these sorts of shifts in reproductive strategy simply reflect the “genetic variation within created kinds”. Why such (repeated) transitions within reptiles is viewed as trivial, but such a transition within mammals would be “notoriously difficult” is not explained.

From egg to placenta

The evidence we see from present-day reptiles thus gives us ideas about how this transition may have occurred in the placental mammal lineage:

1. An egg-laying species shifts towards being able to either lay eggs, or retain them for live birth with a thinned shell, depending on the conditions present. Vitellogenins would be needed and useful at this point.
2. This lineage then becomes increasingly committed to live-bearing, still with intact vitellogenins. Eggs have yolk as per normal, but are not laid any longer.
3. Once committed to live-bearing, embryos receiving nourishment from the uterine wall during the time they are being held becomes advantageous – such as uptake of calcium from the mother, as seen in present-day “dual mode” skinks.
4. Over time, variation and selection allows for the connection between embryos and the uterine wall to become more elaborate – the beginning of a placenta (such as the varied “placentas” we observe in present-day live-bearing reptiles).
5. As a stronger placental connection evolves, there is correspondingly less need for egg yolk to nourish the embryo. Eventually, the vitellogenins are lost, but without impact, since the placental connection has replaced their function.

Of course, these are ideas – hypotheses – about how this transition may have taken place. The point is not that we have determined exactly how it took place, but that it is a probable path based on what we see in other organisms. The point is that this transition seems readily accessible to present-day reptiles – and thus was likely possible for their close mammalian relatives as well. As such, the evidence for vitellogenin pseudogenes we see in placental mammal genomes should come as no surprise.