Does evolutionary theory need a rethink?

[StringJunky]

 

Funny thing is that at least in the plant community, those that are considered leading the field of plant epigenetics (old and new scientists) are typically pretty adamant about how this ultimately ties back to genome evolution (in particular transposons) and is not really environmentally variable. I see a lot more of that come from certain people in the animal community and few of them experts in epigenetics.

 

For years there has been a problem (and continues to be a major problem) that studies looking at DNA methylation or histone modification differences between populations or generations fail to test or control for genetic variation. They make huge claims that they found a DNA methylation difference and therefore its epigenetic and yet not once did they look for any genetic variation that might be causing it. In the cases where genetic variation is also looked at...often its linked to the variation in DNA methylation or histone modification and likely causative.

Confirmation bias would be the potential problem there wouldn't it? If one pursues a line of research that ignores, or doesn’t test for, one of the two possible causes then it’s obvious what the outcome will be.

Edited November 21, 2014 by StringJunky

[chadn737]

Confirmation bias would be the potential problem there wouldn't it?

 

Yes.

 

One of the worst violators of this is Michael Skinner, who has made a lot of news claiming that various herbicides and other chemicals cause transgenerational epigenetic effects.

 

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0102091

 

However, in this study, they use an outbred strain of rats, as opposed to an inbred strain. This means that there is going to be considerable genetic variation floating around from rat to rat and generation to generation. At no point do they test any genetic markers or control for any effects of genetic variation. They test for methylation differences, they find some...which is guaranteed to happen by chance alone, particularly when comparing genetically segregating individuals....and then assert that the effects observed are epigenetic.

 

This is the sort of study one sees time and again, particularly from animal and human community, to a lesser extent in the plant community. Even in some of the better studies that get published in Science or Nature, there is a failure to do this sort of control fro genetic variation.

 

This is exactly the point that Tim Bestor, one of the leading scientists in the field from the animal side has made about many high-profile studies, such that diet of the mother affects the children and subsequent generations.

 

"Other factors could explain the new Science findings, cautions Timothy Bestor, professor of genetics and development at Columbia University. For one, he suggests that the researchers choice of mouse strain could have influenced the study results. Radford and colleagues did not study inbred mice because those mice had more trouble surviving the restrictive diet needed for the experiment. (Inbred mice are often studied because they are genetically homogenous across generations.) Instead, Radford studied more genetically diverse mice that could better survive the poor nutrition regime. In turn, the health issues seen in the second generation of mice (which were born free of the sperm DNA changes seen in the first generation of the mice) could be somehow genetically inherent to this mouse model, he suggests."

 

http://www.scientificamerican.com/article/diet-during-pregnancy-linked-to-diabetes-in-grandchildren/

[starlarvae]

 

No, epigenetics is not central. This is my field. I work on DNA methylation and transgenerational inheritance of DNA methylation in both plants and animals.

 

The overwhelming evidence from both plants and animals is that heritable differences in either DNA methylation or Histone modifications is typically due to genetic variation. For instance, if you find a region of a genome that differs in DNA methylation between two individuals...often times there is some underlying or neighboring genetic variation that is causing this....like a new transposon insertion. In this case, what many people call "epigenetics" actually reduces down to genetics, since DNA methylation and gene expression are phenotypes of a genetic variant.

 

When the variation in DNA methylation or histone modifications are induced by environmental factors, almost never is this heritable, and even then, it is typically stable only for a generation or two. This is not enough time to have long term evolutionary consequences, at least not in anyway to require a rewrite of evolutionary theory.

 

In the field of epigenetics, we do not define it as the environment....gene x environment interactions were known prior to epigenetics and were never part of the original definition. We define it as heritable variation that is not due to underlying differences in the DNA. If that variation is caused by genetic differences...its not epigenetic. If its not heritable and stably inherited...its not really epigenetics.

 

The cases where it is what is called a "pure" epialle (not induced by genetic variation and stably inherited)...these are actually quite rare as a whole and so are the exception, not the rule. We also have little idea how stable they are in the long term. It is also far more likely that if that trait achieves stability, that it is because it becomes a "genetic" trait rather than an epigenetic trait.

 

With the exception of prions, epigenetics operates through the silencing or unsilencing of genes by DNA methylation or histone modifications. While these are common features of development, they are typically tissue specific and reset every generation.

 

Because they are silencing genes, they are actually operating on the DNA already and so ultimately epigenetics is still a function of the DNA. It is hypothesized, and there is some evidence that mutations can accumulate in methylated/silenced genes. Methylated cytosines sometimes become deaminated which can lead to mutation of the cytosine residue. In this way, the silencing of the gene can become encoded genetically through the mutation disrupting that gene function. At this point, the trait becomes genetic.

 

So epigenetics can serve as a transitory phase in evolution, but in the long term, there is no evidence from epigenetic that suggests evolution needs a rewrite. Those who argue that it does, do so not from a standpoint of evidence, but from a standpoint of untested hypothesis and often conviction.

 

Thanks for detailed reply. Please set me straight on something. My understanding is that one surprising result coming out of all the comparative genomics studies is that DNA is highly conserved across species -- that genomic variation across species is not proportional to the phenotypic diversity that we observe across species, but is far less that was expected, based on the vast phenotypic variation that we observe.

 

This disparity, or apparent paradox, (little genetic variation, but lots of phenotypic variation) gets explained in terms of epigenetic regulatory mechanisms. A largely shared set of genes can produce highly diverse phenotypes, in this case, because epigenetic mechanisms regulate when during the life cycle and where in the body various genes are turned on or off. The various regulatory regimes, in turn, determine, to a significant degree, phenotypic differences across species. It's that whole evo-devo thing.

 

If this varying-by-species epigenetic regulation gets turned into genetic differences (somewhere down the line) as you suggest, then why are these epigenetically induced gene differences not more prevalent in the comparative genomics studies? Why does it look like DNA remains highly conserved through evolution if the epigenetic differences get translated into genetic differences? Shouldn't those genetic differences -- to a degree proportional to the phenotypic differences -- show up in the comparative genomics work?

[chadn737]

 

Thanks for detailed reply. Please set me straight on something. My understanding is that one surprising result coming out of all the comparative genomics studies is that DNA is highly conserved across species -- that genomic variation across species is not proportional to the phenotypic diversity that we observe across species, but is far less that was expected, based on the vast phenotypic variation that we observe.

 

This disparity, or apparent paradox, (little genetic variation, but lots of phenotypic variation) gets explained in terms of epigenetic regulatory mechanisms. A largely shared set of genes can produce highly diverse phenotypes, in this case, because epigenetic mechanisms regulate when during the life cycle and where in the body various genes are turned on or off. The various regulatory regimes, in turn, determine, to a significant degree, phenotypic differences across species. It's that whole evo-devo thing.

 

If this varying-by-species epigenetic regulation gets turned into genetic differences (somewhere down the line) as you suggest, then why are these epigenetically induced gene differences not more prevalent in the comparative genomics studies? Why does it look like DNA remains highly conserved through evolution if the epigenetic differences get translated into genetic differences? Shouldn't those genetic differences -- to a degree proportional to the phenotypic differences -- show up in the comparative genomics work?

 

I know of no study that has shown that the genetic diversity is not proportional to the phenotypic diversity. This issue is particularly complicated by the fact that the phenotype is factor of both genetics and environment, i.e. gene x environment interactions. If you have the genetic potential to be 6ft tall but spend your childhood malnourished, you wont be 6ft tall.

 

Furthermore, different variants interact with each other in different ways. This is known as epistasis and this can greatly increase the number of possible combinations and resulting potential phenotypes.

 

Finally, initial studies have missed a lot of genetic variation. SNPs are the easiest to identify via sequencing. There are also chromosomal rearrangements, Copy Number Variants, insertions, deletions, many non-coding variants that have yet to be identified....not to mention that we don't know the extent of rare variants or their effects. Most GWAS studies rely on using known common SNPs and will miss unknown rare variants.

 

So there are many possible reasons for any missing heritability that does not require an invocation of epigenetics.

[starlarvae]

 

I know of no study that has shown that the genetic diversity is not proportional to the phenotypic diversity.

 

Here's a summary of such studies: http://www.scientificamerican.com/article/regulating-evolution/

 

In it the authors Sean B. Carroll, Nicolas Gompel and Benjamin Prudhomme elaborate on the findings coming out of comparative genomics:

 

"For a long time, scientists certainly expected the anatomical differences among animals to be reflected in clear differences among the contents of their genomes. When we compare mammalian genomes such as those of the mouse, rat, dog, human and chimpanzee, however, we see that their respective gene catalogues are remarkably similar. [. . . .] When comparing mouse and human genomes, for example, biologists are able to identify a mouse counterpart of at least 99 percent of all our genes."

"The discovery that body-building proteins are even more alike on average than other proteins was especially intriguing because of the paradox it seemed to pose: animals as different as a mouse and an elephant are shaped by a common set of very similar, functionally indistinguishable body-building proteins."

 

"The preservation of coding sequences over evolutionary time is especially puzzling when one considers the genes involved in body building and body patterning."

 

". . . to our surprise, it has turned out that differences in appearance are deceiving: very different animals have very similar sets of genes."

 

 

Similar observations are presented in this article, http://www.tandfonline.com/doi/abs/10.4161/cc.6.15.4557#preview ,

and here: http://www.nature.com/nature/journal/v402/n6761supp/full/402c41a0.html , in which the author says, "So many examples of [DNA] conservation have now been found that it is no longer considered surprising. We can now state with confidence that most animal phyla possess essentially the same genes, and that some (but not all) of these genes change their developmental roles infrequently in evolution [emphasis added]."

 

So are these guys off the mark? They give the impression that DNA is highly conserved across species. And that phenotypic diversity is due largely to adjustments in gene regulation, particularly involving genes that guide development and the timing of developmental stages, rather than to genetic novelty via mutation.

 

[chadn737]

 

Here's a summary of such studies: http://www.scientificamerican.com/article/regulating-evolution/

 

In it the authors Sean B. Carroll, Nicolas Gompel and Benjamin Prudhomme elaborate on the findings coming out of comparative genomics:

 

"For a long time, scientists certainly expected the anatomical differences among animals to be reflected in clear differences among the contents of their genomes. When we compare mammalian genomes such as those of the mouse, rat, dog, human and chimpanzee, however, we see that their respective gene catalogues are remarkably similar. [. . . .] When comparing mouse and human genomes, for example, biologists are able to identify a mouse counterpart of at least 99 percent of all our genes."

"The discovery that body-building proteins are even more alike on average than other proteins was especially intriguing because of the paradox it seemed to pose: animals as different as a mouse and an elephant are shaped by a common set of very similar, functionally indistinguishable body-building proteins."

 

"The preservation of coding sequences over evolutionary time is especially puzzling when one considers the genes involved in body building and body patterning."

 

". . . to our surprise, it has turned out that differences in appearance are deceiving: very different animals have very similar sets of genes."

 

 

Similar observations are presented in this article, http://www.tandfonline.com/doi/abs/10.4161/cc.6.15.4557#preview ,

and here: http://www.nature.com/nature/journal/v402/n6761supp/full/402c41a0.html , in which the author says, "So many examples of [DNA] conservation have now been found that it is no longer considered surprising. We can now state with confidence that most animal phyla possess essentially the same genes, and that some (but not all) of these genes change their developmental roles infrequently in evolution [emphasis added]."

 

So are these guys off the mark? They give the impression that DNA is highly conserved across species. And that phenotypic diversity is due largely to adjustments in gene regulation, particularly involving genes that guide development and the timing of developmental stages, rather than to genetic novelty via mutation.

 

 

They are talking about differences within protein-coding genes and arguing that much of the differences between species are due to genetic differences in coding regions. This is still genetic variation and requires no rethink of evolutionary theory.

[Endy0816]

Wasn't there something about a greater number of differences in the noncoding regions?

[chadn737]

Wasn't there something about a greater number of differences in the noncoding regions?

 

These are still genetic differences and thus still behave in the ways described by evolutionary and population genetics that underlie the modern synthesis.

[Endy0816]

 

These are still genetic differences and thus still behave in the ways described by evolutionary and population genetics that underlie the modern synthesis.

 

Yes. I mean that between species aren't these regions themselves more varied than the coding sections?

[Robittybob1]

 

Yes. I mean that between species aren't these regions themselves more varied than the coding sections?

Yes because mutations in the non-coding section aren't as often lethal.

[starlarvae]

 

They are talking about differences within protein-coding genes and arguing that much of the differences between species are due to genetic differences in coding regions. This is still genetic variation and requires no rethink of evolutionary theory.

 

I think they're talking about similarities, not differences. Their point is that genomes across species are so strikingly similar, relative to the diversity of phenotypes. Epigenetic gene regulatory mechanisms allow a shared set of genes to produce a variety of phenotypes. This is what happens when cells differentiate in a developing body.

 

Reading up on this topic, I keep coming across the phrase, "non-genetic inheritance," to describe epigenetic inheritance, and even ecological (niche) inheritance. You think that's legit usage?

[Arete]

 

I think they're talking about similarities, not differences. Their point is that genomes across species are so strikingly similar, relative to the diversity of phenotypes. Epigenetic gene regulatory mechanisms allow a shared set of genes to produce a variety of phenotypes. This is what happens when cells differentiate in a developing body.

 

Reading up on this topic, I keep coming across the phrase, "non-genetic inheritance," to describe epigenetic inheritance, and even ecological (niche) inheritance. You think that's legit usage?

 

A regulatory change is not necessarily an epigenetic change, in fact most are heritable changes in regulatory regions and transcription factors.

[Robittybob1]

 

A regulatory change is not necessarily an epigenetic change, in fact most are heritable changes in regulatory regions and transcription factors.

Epigenetics feels like evolution in reserve to me.

[Ophiolite]

Have you reversed some of the letters in reserve?

[Robittybob1]

Have you reversed some of the letters in reserve?

No "reserves" like in a sports team with players on the bench ready at a moments notice. With epigenetics there are genes ready to be switched on when needed.

Edited November 29, 2014 by Robittybob1

[chadn737]

The mechanisms we typically associate with "epigenetics", DNA methylation and certain histone modifications, are most highly concentrated in heterochromatin, in particularly targeting repeats and transposons. It really appears that they evolved with the purpose of preventing the spread of parasitic elements like transposons. Any other function then in actual protein coding genes is more of a co-option after the fact or an accidental spreading from silencing of some of these parasitic elements. Epigenetics is too often thrown around as some mystical magical answer for any hard to explain phenomena...its not.