Search This Blog

A Response to "Questioning Evolutionary Presuppositions about Endogenous Retroviruses"

On an article by Dr. Anjeanette Roberts @ "Reasons to Believe",

My comments are in blue - Barry Desborough, 10 Jan 2016
It is common, in creationist literature, to accuse those who they call "evolutionists" of employing what they call "presuppositions". I find this practise to be utterly intellectually dishonest, in that it tries to deliberately pooh-pooh any conclusions based on evidence and reasoning by branding them so. Here is an article by one such creationist, more articulate than most, where we can examine such an accusation. Roberts' notes are in black, with links in orange.

In a 2006 lecture at Emmanuel College, Cambridge,1 Dr. Graeme Finlay, an immunologist, cancer biologist, and Christian, made some remarkable observations about the genetic similarities of human and nonhuman primate (NHP)2 genomes. He drew the conclusion that these similarities presented incontrovertible genetic evidence for the common ancestry of humans and other primates.3 In doing so, Dr. Finlay employed his expertise in cancer biology to explain the evidence for common descent. I found his presentation compelling.

I would urge the reader to read Finlay for themselves, rather than rely on the say-so of a creationist. He has produced an excellent book about genetic markers, including ERVs, that shows why common ancestry is the only viable account of the evidence available. 

Much of what goes into scientific hypothesis formation and experimental design is based on assumptions and facts resulting from previous observations and studies. The orderly and reliable workings of nature and the reproducibility of outcomes in controlled scientific experimentation is at the heart of scientific inquiry and technological advancement. Without reasoned inference from previous observations, scientific advancement would be paralyzed.

I have one quibble here. "Based on assumptions". There is a place for making assumptions in science, but they must be provisional, working assumptions that are to be rejected as and when any observations falsify them. To build on previous work, new enquiries must have confidence  in only those "assumptions" that have been shown to be reliable - thus they are no longer "assumptions", but "conclusions" and "findings".
Applying inductive reasoning to comparative genomic analyses is one way scientists search for functional elements within various genomes. A scientist assumes that if the function of a given sequence within a given genome is known and if he finds an identical or similar sequence in another organism’s genome, then he can provisionally assume a similar (if not identical) function in the second organism. Of course, he cannot prove the same function exists in the unknown genome until he designs experiments with proper controls that allow him to actually measure and compare the functions of the similar sequences. But often, especially for multicellular organisms, such an experiment is too complex. So he relies on inductive reasoning to reach some conclusions. This is a key concept in many genomic studies and in Dr. Finlay’s assertion of genetic evidence for common descent.

Let us see if this assertion is so as we proceed.
Virus-Infected Cells, Clonal Expansion, and Common Descent
In his lecture, Dr. Finlay explained how the infection of cells with a type of retrovirus—human T-cell leukemia virus (HTLV)—begins with multiple random insertion events of the HTLV proviral DNA into various sites of the host cell’s chromosomal DNA. Much later, after infection with HTLV, leukemia may develop. When it does, it is observed that every leukemic cell within the infected individual shares the exact same insertion site for the HTLV proviral DNA.4 When a retrovirus infects a cell its genomic sequence is inserted into the host cell’s chromosome. So a common shared site of insertion in every leukemic cell indicates that the leukemia results from clonal expansion of one single, HTLV-infected cell and not from multiple independently infected cells. If leukemia arose from multiple independently infected cells, the proviral insertion sites would differ among the host’s leukemic cells. Leukemia, therefore, results from one unique insertion event (among many). This clonal expansion of rogue cells is a well-established hallmark of cancers.

So this is no presupposition, but a conclusion drawn from the evidence, (surveys of actual integration sites) that retroviruses do not target specific integration sites. The only mechanism that can produce the result of identical sites is reproduction from an single original cell. 
Dr. Finlay took this well-established fact of clonal expansion and applied it to comparative analyses of chromosomal sequences of humans and NHPs. The chromosomes of all hominids are riddled with sequences known as endogenous retroviral (ERV) elements. About 8 percent of the human chromosome is composed of ERV sequences of unknown etiology (origin). ERVs are so named because they share sequence homology and other characteristics with known retroviruses. But unlike modern retroviruses (such as HIV), no human ERV is still functionally infectious. Not only are the genomes riddled with ERVs, but one also finds shared ERV sequences at specific identical insertion sites in the various hominid chromosomes. Since ERVs are no longer infectious and yet occur in identical places in the genomes of hominids, by inference, according to the clonal expansion theory, they are presumably indicative of descent from common forebears. Put another way, for all humans to possess the same set of 400,000 ERV elements “as part of their common genetic endowment, the germ-line cells sustaining the original infections must have been ancestral to us all.”5

Here, Roberts slips in the assertion that ERVs are of unknown origins, and that the notion that they are thought to be related to known retroviruses because of shared sequence homology (similarity of form). There is much more evidence for the relation than she is letting on here, and I will go into it in due course.
And for humans and NHPs to share several thousand chromosomal ERV insertion sites this must reflect a common ancestor that was infected with an ERV progenitor before the various hominids diverged evolutionarily. Like leukemic cells, although the initial viral insertion events were random, the fact that all descendants share identical insertions in their genomes indicates proof of a single infection and insertion event and, therefore, common descent.
Examining Presuppositions
Dr. Finlay asserted this as staggering and incontrovertible evidence for common descent; and taken with his presuppositions, he is absolutely correct. However, his presuppositions are many and, although plausible, they may not be as solid as he or other neo-Darwinian evolutionists think.
His presuppositions follow this line of reasoning: (1) ERVs are evolutionary hallmarks of viral infections and insertion events; (2) infection by ERV viral progenitors occurred in gametes or gametic precursors of a common ancestor of hominids; and (3) following infection, ERV progenitor viral sequences were inserted randomly into the genomes of these ancestors and were subsequently passed on to all descendants, including all divergent species that branched off from the ancestors who incurred the initial infectious events. Additionally, he presumes that ERVs are by-and-large nonfunctional within the genomes in which they are found today. As a molecular biologist and virologist I take exception to a few things and think there is room for questioning Dr. Finlay’s presuppositions.

So here we go. Let's see if these accusations of working from "presuppositions" are warranted.
In addressing the first two presuppositions, consider the following. Virus replication is known to be restricted on one level by the types of cells any given virus can infect. This restriction is often mediated at the level of viral entry into the cell by protein receptors on the cell surface. Human retroviruses are known to infect somatic cells, primarily of hematopoietic origin, not gametes or gametic precursors. Gametic cell-types apparently lack the appropriate receptors for viral entry.

Retroviral replication is known to be highly error-prone. Mutations are not subject to any correction processes. The hugely parallel replication process will throw up large numbers of mutants - mutants that can cross over to different species, let alone different cell types within a species. The assertion that "gametic cell-types apparently lack the appropriate receptors for viral entry" is, I am sorry to say a *cough* presupposition. In fact, we know it is a false presupposition. It takes just one observation to falsify it. KoRV. Not only is there evidence that it has crossed over from another species, but there are exogenous and endogenous versions of it. How is it supposed to have entered the germ-line other than by endogenization?

How is it justifiable to call conclusions based on this sort of evidence and reasoning "presuppositions"?
A second obstacle to proviral insertion is raised by the observation that retroviral DNAs of some retroviruses do not integrate into the DNA of quiescent (inactive), nonreplicating cells7—such as gametes.8 Therefore, proviral integration into an ancient gametic chromosome faces two known obstacles: one at viral entry and another at proviral integration. These are nontrivial scientific objections to Dr. Finlay’s first two presuppositions. Nevertheless, if ancient retroviruses were able to infect the gametic precursors of a proposed common ancestor, then the existence of shared ERV sequences at identical sites in all (or almost all) descendants’ genomes would be expected.

Although the assertion that "some retroviruses do not integrate into quiescent cells" contains a reference (7), there is no study cited in note 7.  It is merely another assertion. 

Debate continues whether retroviral infection and integration occur in quiescent cells. It is an extremely infrequent event if it occurs. If it does occur insertion is less random and appears to be preferentially directed to regions of actively transcribed genes.

It is true that some retroviruses show a statistical tendency to integrate in active regions, it is a mere statistical tendency. Stuff may happen less frequently, but it still happens. In accusing Finlay of presupposition here, Roberts fails to take into account all the other items of evidence that, as we shall see later, make endogenization a sound conclusion.
Although tantalizing observations of the koala retrovirus (KoRV) activity in captive and wild koalas may shed light on elements of endogenization (establishment of proviral sequences in the germ line), we should resist inferring too much from these observations.9 Recent emergence (estimated date circa 1900) of KoRV may or may not render relevant insight into the presence of shared ERVs in hominid genomes. Like most other retroviruses and ERVs, the origin of KoRV and the closely related gibbon ape leukemia virus (GALV) found in captive gibbons is unknown. These retroviruses utilize a class of cellular receptors for viral entry which results in increased host range (primate and marsupial) and broader cellular tropism (somatic and germ line cells in koalas).10 The greater distribution and shared similarity of the cellular receptors within and across species may be a key factor for the endogenization being observed and, therefore, one must not infer too much from limited data.

KoRv is but one line of evidence that leads us to the conclusion that endogenous retroviruses are produced by exogenous viruses integrating into germ-line cells. But it is a powerful one. Roberts fails to justify the assertion that we should resist perfectly reasonable inferences  from this.
Insertion Sites Are Not “Random”
Despite early findings in vitro, retroviral insertion sites are not selected randomly. Various retroviruses have varying degrees of insertion site preferences, some showing site bias and others even demonstrating integration specificity at the primary sequence level. There are a variety of factors now known to effect integration site specificity.11These include different viral proteins (IN, Gag, U3 LTR), chromatin accessibility (A/T-rich distorted DNA and outwardly facing major grooves), cell-cycle effects (integration in dividing cells occurs at a much higher rate, and increased site specificity is observed in integration in non-dividing cells), and cellular integration co-factors (tethering proteins like LEDGF/p75, gene regulatory elements, and epigenetic marks). Since a range of insertion site specificities and contributing factors exist for the various classes of retroviruses, it is possible that retroviral infections establishing the shared ERV sites in NHP chromosomal segments had even greater specificity for insertion site selection than those actively tested and observed to date.

Nobody says that retroviral integration is entirely random in terms of integration site loci. Various retroviruses do indeed have varying degrees of insertion site preferences, but this is a statistical phenomenon. Surveys of actual integration sites in the same cell types in the same organism show great variation in integration site locus. A statistical tendency to integrate in certain types of regions is nowhere near specific enough to account for common loci in different species. I can't believe that Roberts does not know this. See

Speculation about greater specificity in the past is pure speculation, without any justification. Indeed, the resurrection of a long inactivated retrovirus ("Phoenix") shows no greater specificity. Integration is controlled by the retroviral enzyme integrase, that is not greatly different from one ERV or one exogenous retrovirus to another.
Furthermore, despite heritability of ERVs at shared insertion sites, absence of specific ERV sequences in some NHP genomes challenges the common descent paradigm. HERV-K GC1 is found in chimps, bonobos, and gorillas, but not in humans.12 PtERV1 is present in chimps and great apes, but not in humans and orangutans.13 As others have noted, these findings undermine the notion that an ancient infection invaded an ancestral primate lineage, since according to phylogenetic analysis of species, great apes (including humans) share a common ancestor with Old World monkeys.14 Another observation of shared NHP ERVs that is contrary to evolutionary predictions involves the divergence of sequences in paired 5’ and 3’ proviral LTRs,15 which accrue differing mutations at similar rates following insertion. Divergence between LTR sequences at a single shared ERV site sometimes varies more significantly in one species than in another, suggesting differences in times of insertion between the two species.16 For example, estimated divergence of chimpanzee 1p31.1a proviral LTRs is 6.5 times greater than that observed in humans. Human 1p31.1a is also dimorphic with solo LTR and provirus, unlike the chimp ortholog. Both of these findings suggest a much more recent integration event in humans than in chimps at orthologous sites.

Re. PtERV1, these are not in common loci in chimps and other great apes, so they are neither evidence for common ancestry, nor are they evidence against. They are independent endogenizations.

Re. differing LTR divergence between orthologous pairs of LTRs, Roberts does not provide a reference, but differing fitness contributions and therefore differing purifying selection levels could well account for it. Age estimates are based on surveys of large numbers of data. Cherry picking individual outliers is inaccurate, and if it is done deliberately, it is dishonest. 

Update, 12 Jan 2016: As Kenneth Gilmore points out in his blog, the "LTR" study that Roberts appears to be referring to is Genome-wide amplification of proviral sequences reveals new polymorphic HERV-K(HML-2) proviruses in humans and chimpanzees that are absent from genome assemblies by Catriona M Macfarlane and Richard M Badge. Gilmore points out that the paper discusses sequence homogenisation as an explanation for the incongruous result. Roberts' use of this "example" appears to be a deliberate attempt to mislead.
Although evolutionary arguments are made to account for these observations,18independent ERV infection events with similar insertion site specificities offer simpler viable explanations for ERVs that do not track with phylogenetic predictions based on NHP species relatedness. The Reasons to Believe model of common design also offers a simpler explanation for the possibility of multiple ERVs that do not follow primate phylogenetic trees: They serve relevant functions and are not evolutionary artifacts.

There are no "ERVs that do not track with phylogenetic predictions based on NHP species relatedness"!

Re. function, see
Important Sidebar on NHP Genome Analyses
Evolution proponents commonly presuppose that ERVs—along with most other repetitive elements (REs) found in the human genome—are nonfunctional markers of evolutionary processes. In adopting this posture many believe we have sure knowledge for such inductive conclusions. They appeal to whole genome sequencing (WGS) data of human, chimp, gorilla, bonobo, orangutan, Neanderthal, Denisovan, and many others. From these sequences, they highlight data that support their presuppositions and construct various scenarios to account for data that go directly against their predicted model.

Non-creationists do NOT presuppose that ERVs are nonfunctional! In fact, it is real scientists who have found that certain components of certain ERVs have function. See the link immediately above. This is a common, intellectually dishonest creationist tactic: create a strawman, attack it, and declare that you have scored a "point". Creationists are driven to such desperate lengths because they cannot answer the actual case from ERVs.

We have yet to see any data that go "directly against their predicted model."
These WGS accomplishments are truly remarkable; however we fail to acknowledge that vast troves of information are still uninterpretable and inaccessible. Although it is widely accepted that the human genome sequence was “completed” in 2004, to this day nearly 10 percent of it remains unsequenced because it exists in inaccessible, densely packed heterochromatin (~8 percent) or is replete with repetitive sequences that are not yet possible to assemble (~2 percent).19 The majority of these sequences are simply omitted from genomic comparisons for three reasons:
  1. The tight association of (constitutive) heterochromatin with its associated proteins makes it inaccessible to current methodologies employed in retrieving chromosomal DNA for analysis.
  2. Heterochromatin is replete with highly repetitive sequences, as is much of the euchromatin sequenced to date. It is technically very difficult to sequence and accurately align these types of highly repetitive sequences because of the very short “reads” and currently available alignment algorithms.
  3. The lack of intercellular regulatory access to much of the tightly bound heterochromatin and the basic characteristics of repetitive elements have left many evolutionary theorists arguing for their relative insignificance in function and comparative genomic studies.
All of these arguments rest primarily on mistaken inferences that these sequences (those inaccessible in heterochromatin and the highly repetitive sequences in euchromatin) are of little to no significance. A critical flaw in this presupposition is a failure to recognize humbly that our efforts at unraveling the complexities of the human genome are still in their infancy.So - we don't know everything, so we can't know anything? I wonder how many creationists would agree with this statement, when applied to their own beliefs?
As with insertion site randomness, more recent scientific findings challenge each of these presuppositions. Genes and regulatory elements have been discovered in heterochromatin. Heterochromatin sequences may in fact have specific developmental (or tissue) roles not captured in snapshots of genomic analyses, which are sometimes performed on immortalized cell-lines rather than on primary, tissue-specific cells.21Along these lines, information within published genomes may limit our understanding of differences that occur within various tissues, in various stages of development, and under various stresses. Although these variations are unlikely to drastically change primary DNA sequences (unless copy number variations are found to have more significant roles in phenotypic variability at the cellular or organismic level than currently understood), they could certainly contribute to a change in chromosomal architecture and activity. That which is envisioned as fixed may in fact be revealed as having fluidity capable of significant impacts at an individual level. Data collected in the future from transcriptome (RNA molecules) and proteome analyses will help define differences—but with this added data the information to be sorted will grow exponentially.

And this is relevant - how?
ERVs and REs Have Significant Function
Researchers have found that REs (like ERVs), once thought to be junk, can be functional (see here and here).22 Some ERVs are transcribed and have specific functions in various cells and tissues. Some REs provide regulatory functions themselves and some affect the proximity of other regulators to specific genes, which also affects expression and function. Additionally REs often occur in long stretches along chromosomes that may provide fluidity for genomic restructuring in complex adaptive situations. In light of these discoveries, it is important to emphasize that additional functions are bound to lie hidden in previously identified ERVs and in ERVs not yet identified.

See ERVs promote the transcription of host DNA. Doesn't this prove they are designed?
According to RTB’s creation model, those ERVs or REs held in common with minimal divergence with other NHPs are likely to serve roles common to all. ERVs or REs that differ significantly or are distinctive to humans (or even exclusive to some particular humans) likely contribute to cross-species and interspecies unique attributes.

A neologism comes to mind "metooism". 
Other considerations that should not be ruled out in forming presuppositions regarding the presence of shared ERVs in primate genomes depend on the fact that the origins of retroviruses and other viral families are unknown. All viruses depend on living cells in order to replicate. A virus cannot replicate independent from a living cell, which supplies it with energy and mechanisms for replication. The RTB model proposes a Creator who designed with foresight and finesse adaptive mechanisms that would provide organisms the ability to persist and even thrive in conditions of change and stress. Since scientists have discovered function for some ERVs, it is possible that ERVs and transposable elements may be part of these mechanisms.

How does one "form a presupposition"? I thought that presuppositionalism precluded "formation". I thought it was like an anvil, falling upon one from a ten-storey building. The origins of retroviruses is completely irrelevant to the case for common ancestry from ERVs. They could have arisen naturally, been created by aliens or by supernatural entities, either good or evil. The fact remains that, whatever their origins, they left their traces in our genomes, testifying to our common ancestry with other species.
Evolutionary Presuppositions Need to Change
In light of increasing evidence challenging evolutionary presuppositions, their staying power is quite unfortunate. The most detrimental aspect of accepting the evolutionary explanation of ERVs in human and NHP genomes is that it inhibits scientific inquiry and progress by attributing no other significance to REs and ERV-like elements maintained in NHP genomes than that of evolutionary artifacts. Forcing sequences into a paradigm that renders them insignificant and useful only as evolutionary markers stifles us from probing these sequences for unique, shared, or distinguishing functions. It is in a sense what others have called a classic example of derailing an objective analysis of the data.

This is patently false, and unconscionable. There is intense activity in real science, studying retroviruses and ERVs. They provide potential ammunition in the fight against disease, and they shine a light on our evolutionary past. All of the scientific research that creationists cherry-pick, misrepresent and spin has been done by actual scientists, dedicated to the truth as opposed to some presupposed narrative. The irony is about as deep as deep can get.
We need to remember (or consider for the first time) that we’ve only just begun to unravel the human genome and all its complexities. It will take decades of dedicated, well-designed, nonpresumptuous research to unpack it. It is no surprise that, as we do, the intricacies and complexities we discover will stretch our imaginations. The intricate networks demonstrate extraordinary orchestration and fine-tuning, which point to incredible, complex design. The greater the complexity and functionality of the human genome, the greater the impedance to plausible neo-Darwinian explanations. Yet discoveries like these, pointing to incredibly intricate and complex designs, are exactly what the RTB creation model predicts. It is surprising that the more we learn about the magnificent orchestration of the diverse symphonies playing inside each cell, the fewer people recognize the obvious themes of a brilliant Composer.

Yes "Evolution is cleverer than you are." - Orgel's Second Rule.

Go to
my FAQ for a summary of the case for common ancestry from commonly located ERVs and the evidence for it.

(Return to the ERV FAQ)