Feather Genes Surprisingly Early – But Misinterpreted

An interesting paper…

Feather development genes and associated regulatory innovation predate the origin of Dinosauria by Craig B. Lowe, Julia A. Clarke, Allan J. Baker, David Haussler, Scott V. Edwards.
MBE Advance Access published November 18, 2014 Society for Molecular Biology and Evolution. Letter to MBE – “Discoveries”.

…recently threw new light on how feathers evolved – though the story isn’t yet perfectly clear!

The evidence they found can paint a different picture, and as my previous blog posting shows, my views seem to be the ones that turn out right.

This study found that the machinery for making feathers seems to have evolved earlier than we had thought – or at least earlier than the experimenters had thought. I knew they’d be earlier than that, but I didn’t think the genetic traces would go back this far!…

[Shouldn’t be any propblems downloading and saving this webpage now but there’s also this pdf version.]

Other studies had concentrated on stretches of DNA that defined proteins, but this study also looked at pieces of DNA, 600,000 of them, that didn’t. They are often positioned next door to the protein-coding parts, but instead of describing the structure of proteins, they control when and in which tissues the proteins should be knitted. Evolution hadn’t allowed these Conserved Regulatory Elements to be randomised or lost so they seem to be important, and as they don’t describe proteins, we assume thay must help control their production.

These Conserved Regulatory Elements are also known as Conserved non-exonic Elements (CNEEs), because pieces ofDNA that translate into protein are called exons. (This was an unfortunate historical naming accident, as I explain in my book! (US/UK), or see top right.)

In the RIGHT half (b) we see a big green peak.  It shows when the genes for making the β-keratin protein (which makes up most of a feather in its final form), approached 20 times the expected average rate of evolution: somewhere on the line to birds but after crocs.  The tiny green peak just to the left of the big peak, shows a burst in activity in these genes that even turtles shared (with birds and crocs).  Something, prior to turtles, needed a special supply of the stiffest keratin!  My account explains this; the authors' doesn't. In the LEFT half (a), the green line here shows the rate of evolutionary activity of the genes needed for making the proteins needed for making a feather apart from β-keratin.  It had flat-lined by the time it got to archosauria, which means the definition was complete of non-β-keratin proteins needed for making feathers, even before the last common ancestor of birds and turtles (inc. crocs).   The blue line in the left half shows rate of evolution in the genetic machinery for controlling the expression of proteins shown by the green line: one peak is prior to the arrival of what non-biologists call reptiles, called Amniota here ("Reptilia" at the foot of the chart refers to higher animals not including things like mammals).  A second peak at Archosauria shows that some further rearranging of the expression of whatever feathers were then, was needed between the evolution of turtles and crocs.  Note: the graphs in the left half (a), both blue and green, don't go above about 1.7 according to the scale down the left, whereas even the tiny green blip in the right image (b) goes up to 2.  At that last little kink in the green line on the left, before it flat-lines, it's about 0.24, which is about 14% of the highest point in the green line (about 1.7), meaning that 86% (100%-14%) of the work of the green line had been done prior to "Reptilia", and so must also therefore have been prior to Dinosauria.

In the RIGHT half (b) we see a big green peak. It shows when the genes for making the β-keratin protein (which makes up most of a feather in its final form), approached 20 times the expected average rate of evolution: somewhere on the line to birds but after crocs. The tiny green peak just to the left of the big peak, shows a burst in activity in these genes that even turtles shared (with birds and crocs). Something, prior to turtles, needed a special supply of the stiffest keratin! My account explains this; the authors’ doesn’t.
In the LEFT half (a), the green line here shows the rate of evolutionary activity of the genes needed for making the proteins needed for making a feather apart from β-keratin. It had flat-lined by the time it got to archosauria, which means the definition was complete of non-β-keratin proteins needed for making feathers, even before the last common ancestor of birds and turtles (inc. crocs).
The blue line in the left half shows rate of evolution in the genetic machinery for controlling the expression of proteins shown by the green line: one peak is prior to the arrival of what non-biologists call reptiles, called Amniota here (“Reptilia” at the foot of the chart refers to higher animals not including things like mammals). A second peak at Archosauria shows that some further rearranging of the expression of whatever feathers were then, was needed between the evolution of turtles and crocs.
Note: the graphs in the left half (a), both blue and green, don’t go above about 1.7 according to the scale down the left, whereas even the tiny green blip in the right image (b) goes up to 2.
At that last little kink in the green line on the left, before it flat-lines, it’s about 0.24, which is about 14% of the highest point in the green line (about 1.7), meaning that 86% (100%-14%) of the work of the green line had been done prior to “Reptilia”, and so must also therefore have been prior to Dinosauria.


These controlling elements seem to have changed fastest at two times in evolutionary history: first, when amphibians were evolving into things that laid dry-shelled eggs, and second, when some reptiles were giving rise to the archosaurs – the group that contains the most spectacular reptiles such as birds, pterosaurs, crocodiles, dinosaurs etc.

86% of the regulatory elements (known as CNEEs) for birds’ feathers other than those controlling beta-keratin, are already present in crocs; the authors make the fairly reasonable assumption that they were already present in the ancestor of birds, crocs and dinosaurs. Genes that birds have changed on their own include IGFBP2 and IGFBP5, which affect body size amongst other things. The authors believe, wrongly in my view, that birds’ ancestors shrank rapidly in size just before they started flying, and that they have evidence for it here:

“We estimate that 86% of such [feather?] regulatory elements were present prior to the origin of Dinosauria. On the branch leading to modern birds, we detect a strong signal of regulatory innovation near IGFBP2 and IGFBP5, which have roles in body size reduction, and may represent a genomic signature for the miniaturization of dinosaurian body size preceding the origin of flight.”

…or then again, this “miniaturisation” stage might not even have existed. (See later.) [It’s not clear to me whether that “86%” applies to all CNEEs, including some that control body size, or just to feather stuff 🙂 .]

“Although feathers were long thought to be a key innovation associated with the origin of avian flight, paleontological discoveries over the past fifteen years indicate a more ancient origin; filamentous feather precursors are now known to be present in many lineages of non-avian dinosaurs, and pennaceous feathers clearly arose prior to the origin of flight [they give an unbalanced list of refs].”

Actually pennaceous feathers clearly arose prior to the authors’ idea of the origin of flight (in their view, powered flight evolved with no preceding gliding stage). Another hypothesis, where gliding flight started at the same time as an important phase of innovation which they report for the main feather protein, therefore explains the evidence better, but is ignored by the authors because they think they can get away with it. One theory should get a green light over another if it explains the evidence better, and in palaeontology, the light ain’t gonna get any greener than this, for the theory they not only disfavour but pretend doesn’t even exist.

The same information is here presented in a different form, with the three lines in the backbone of the tree showing the amount of evolutionary change happening at (or before!) various stages, with red meaning a lot of evolutionary change and blue, little. The bottom line means genes for coding specifically beta-keratin; the middle line ("patterning genes") means genes coding for non-beta-keratin proteins used for making feathers, and the CNEEs line refers either to genes for controlling the making of feathers or controlling the expression of anything (maybe the latter). The lower, "keratin genes" line did little to enhance the basic keratin ability, until just prior to turtles (prior to Archosauria proper), by which time very primitive gliding apparatus was needed. (Turtles are turtle-like today but not all their ancestors were!) Between crocs and birds, plenty of improvement was needed on the keratin genes, shown, on that lower line, by a massive change after crocs and before birds (remember the big green peak in the right half of the first image). The top line seems to show the supposed pre-bird “shrinkage” as no more than an average degree of mutation not ascribable to anything particular. Nice work getting the data but... well I don't use the word speculation in criticisms but if I did I'd use it on their claim that this is evidence for shrinkage. Instead I'd say their theory didn't explain the observations better than any other. However everything springing from the backbone of the chart from just before the point where turtles split off, shows evidence for a feather inheritance, so the best theory is that they'd had feathers: initially not fibrous; initially not much on the arms. That's why the joint ancestor of birds and turtles needed to re-arrange its beta-keratin machinery (not because the ancestors of birds needed turtle-like back scutes!). Nothing on the spine of the tree looked more like a dinosaur than a bird.

The same information is here presented in a different form, with the three lines in the backbone of the
tree showing the amount of evolutionary change happening at (or before!) various stages, with red
meaning a lot of evolutionary change and blue, little. The bottom line means genes for coding
specifically beta-keratin; the middle line (“patterning genes”) means genes coding for non-beta-keratin
proteins used for making feathers, and the CNEEs line refers either to genes for controlling the making
of feathers or controlling the expression of anything (maybe the latter). The lower, “keratin genes” line
did little to enhance the basic keratin ability, until just prior to turtles (prior to Archosauria proper), by
which time very primitive gliding apparatus was needed. (Turtles are turtle-like today but not all their
ancestors were!) Between crocs and birds, plenty of improvement was needed on the keratin genes,
shown, on that lower line, by a massive change after crocs and before birds (remember the big green
peak in the right half of the first image). The top line seems to show the supposed pre-bird “shrinkage”
as no more than an average degree of mutation not ascribable to anything particular. Nice work getting
the data but… well I don’t use the word speculation in criticisms but if I did I’d use it on their claim that
this is evidence for shrinkage. Instead I’d say their theory didn’t explain the observations better than
any other. However everything springing from the backbone of the chart from just before the point
where turtles split off, shows evidence for a feather inheritance, so the best theory is that they’d had
feathers: initially not fibrous; initially not much on the arms. That’s why the joint ancestor of birds and
turtles needed to re-arrange its beta-keratin machinery (not because the ancestors of birds needed
turtle-like back scutes!). Nothing on the spine of the tree looked more like a dinosaur than a bird.

In the DNA, they found 13,307 CNEEs (of that original 600,000) that seemed to be associated with the 193 apparently feather-related genes they had found (largely because they were close by). Out of those 193 apparently feather-related genes, 67 seemed to be involved with making the keratin protein that makes up the finished feather, and 126 seemed to do something else. These groups were examined separately.

The DNA in the 67 genes defining the keratin protein molecule underwent very rapid evolution at about the time, the authors claim, the first birds seem to have evolved – actually they mean at the time the first dinobirds with powered flight appear in the fossil record. That time, if it was at that time, which it probably wasn’t (see below), was actually the time specifically when individual feathers first seem to have been used on their own for thrust generation, in particular in Archaeopteryx. Its apparent ancestor Anchiornis seemed to use the whole end of its wing for generating lift and perhaps thrust, as some birds tend to today (falcons, gulls, glossy ibis etc etc.)

Most people will be as surprised as the authors were that the genes responsible for proteins other than keratin used in feathers seem to have almost finished their evolution at the time reptiles, never mind archosaurs, were first appearing, and had stopped by the time the archosaurs did appear. This means that pretty well all the types of proteins used in our skin and those of the other animals above amphibians, had been perfected by the time reptiles arrived. After that, making bare skin or feathers, hair or any kind of “scale”, was just a matter of re-arranging the standard bricks in different patterns.

Notable is the choice the authors refreshingly explicitly express: that in the archosaurs, filamentous or bristle structures either originated once early in the clade or three or more times. (Actually the filamentous structures probably did evolve more than once – it was the original blade-form protofeather that originated once):

“Thus, the genic [for making the molecules] and regulatory [for organising them] complement [bits of DNA] identified in the ancestral archosaur was either a flexible toolkit coopted in multiple origins of new structures including feathers, or indicates an ancient origin in that clade for filamentous integumentary structures, often called feather precursors, on some part of the body or stage in development more than 100 million years before the origin of pinnate feathers in dinosaurs.”

[Smart money in my bold.]
There was a flexible toolkit for skin structures in general, as we’ve already seen, but feathers were evolved once and then multiply-inherited (and often wildly transformed). Here are some of Tracy Ford’s illustrations showing non-bird adaptations of protofeathers, appearing in The Secret Dinobird Story US/UK:

fig64
fig63
It’s odd, the way keratin (the beta-keratin used in feathers) appears to have gone through just two expansion stages. Remembering that we inherited basic keratin from way back even before recognisable fish, and that even lizards and therefore presumably all modern reptiles have some beta-keratin, the first burst of its evolutionary activity after that seems to have been just in time for its fruits to be enjoyed by the common ancestor of turtles and birds. In my scheme, that corresponds to gliding (on pre-feathers springing from the spine) in the precursors of the archosaurs, if turtles aren’t quite in archosaurs, and leaves its physical echoes not just in the feathers of birds but, much adapted, in the plates of tortoises and turtles. This first burst of genes coding for beta-keratin, though small in comparison to the second, is actually quite big compared to the activity levels discovered in non-beta-keratin genes, reported above.

The second burst of beta-keratin-coding gene evolution was massive and occurred somewhere between the first archosaurs and birds… by which is meant modern birds, though it seems to me it might have occurred any time from the last ancestor of birds and crocs, until the last ancestor of all modern birds – an interval > 100,000,000 years, and in no way pin-pointable to the late Jurassic. But that big second burst is entirely consistent with either the theory that feathers were first used for some kind of flight just before Archaeopteryx, or alternatively, way back in the Triassic, probably for gliding, but now needing feathers on the limbs. However the first burst, in the turtle-bird ancestor, is better explained by that ancestor needing a bit of special flight-quality keratin for something (e.g. gliding on feathers down the spine).

The authors note a somewhat surprising lack of evolutionary novelty in genes coding not for β-keratin itself but for its control. However, as I’m sure they realise, this is just a repeat of the earlier picture where the basic machinery for creating skin was put in place very early and just needed a bit of tweaking to distinguish between the major lineages within the tetrapods.

They also comment on a burst of gene activity that seems to correspond to all sorts of body growth activity in modern birds (mentioned a couple of times above), that occurred “prior to their appearance” (i.e. after the split with crocs but before the LCA of moderns). They say it relates to limb growth and other aspects of growth, and is to be expected. However, they also hope to suggest it is evidence for a shrinkage in the size of birds “after they evolved from dinosaurs”. Unfortunately, it is also consistent with all the development birds needed from the time they split from the croc lineage, and as birds are pretty different from crocs in a large number of respects, we don’t necessarily need to point definitely to a process of shrinkage. In fact the arms will have expanded, and the evolution in body format is also consistent with development to birds from a croc-bird ancestor no bigger than modern flying birds. There is no special evidence here for shrinkage full stop, never mind shrinkage just prior to Archaeopteryx.

The paper assumes that whatever evidence may appear through this investigation or any other, birds descended from dinosaurs, and no other possibility will ever be considered. (As I explain at length in my book (US/UK) to those willing to follow the science carefully, this is completely unjustified.)

That change in genes which control body and limb size and format will include tens of millions of years from the early Triassic, when they were small, up to about the middle of the Cretaceous, by which time they will have gone through the volant dromaeosaur stage, but might still have been crow-size. But that will have involved large-scale changes to the rear limbs and tail, followed by larger changes in the opposite direction when they went from the volant dromaeosaur (vodrom) bauplan to that of the modern bird. The hip structure was massively changed.

(Subsequently – and this is beyond the second big burst – modern birds and their genes would have had to have changed from whatever form they were in when they survived from the Cretaceous into the Tertiary, to today’s form. The modern chickens or turkeys (whose DNA was implicated in this investigation) will not have been in that form at the K-T boundary. I’m guessing “chicken Antarctica” would not have been a miraculous survival ticket, though in fairness that might not have been too different from the form the ratite ancestors survived in; however I would envisage more duck-like. The zebra finches used in the study will have needed to go from their K-T transition form (I like to imagine heron-like but water-bird is as much as we can be fairly sure of), to small water bird, to bird of prey and eventually to sparrow-form. They did go through the bird-of-prey stage, as I alone predicted in 2008, according to Jarvis et al. 2014 “Whole-genome analyses resolve early branches in the tree of life of modern birds.”Science 12 December 2014, VOL 346 ISSUE 6215.)

Their final graphic shows parts of the genome where DNA which apparently controls gene expression, has changed a lot on the lead up to birds. Near these places, shown as red spikes or rectangles, or hollow green boxes, are shown the parts of the bird whose development seems to be controlled. The top third and the left half of the middle third, show the full genome. This genome has been put together as a generalised archosaur genome, made up from turkey, rock pigeon, zebra finch etc., listed in the bottom left hand corner. The right half of the midddle line (b) is a close-up of the right-hand end of the top line, and the bottom third is in turn a close-up of part of b. These close-ups show how close two genes, IGFBP2, IGFBP5, apparently linked to “body” growth, are, to controlling genes, CNEEs (especially red boxes) that underwent heavy evolutionary change in the > 100 million years prior to the last common ancestor of modern birds. It is interesting, but it doesn’t support their theory that dinosaurs shrank prior to the emergence of birds with powered flight, more than it does anyone else’s, since everyone’s theory predicts massive change between the last common ancestor of crocs and birds, and the LCA of modern birds.

Their final graphic shows parts of the genome where DNA which apparently controls gene expression,
has changed a lot on the lead up to birds. Near these places, shown as red spikes or rectangles, or
hollow green boxes, are shown the parts of the bird whose development seems to be controlled.
The top third and the left half of the middle third, show the full genome. This genome has been put
together as a generalised archosaur genome, made up from turkey, rock pigeon, zebra finch etc., listed
in the bottom left hand corner. The right half of the midddle line (b) is a close-up of the right-hand end of
the top line, and the bottom third is in turn a close-up of part of b. These close-ups show how close two
genes, IGFBP2, IGFBP5, apparently linked to “body” growth, are, to controlling genes, CNEEs
(especially red boxes) that underwent heavy evolutionary change in the > 100 million years prior to the
last common ancestor of modern birds. It is interesting, but it doesn’t support their theory that dinosaurs
shrank prior to the emergence of birds with powered flight, more than it does anyone else’s, since
everyone’s theory predicts massive change between the last common ancestor of crocs and birds, and
the LCA of modern birds.

Advertisements
This entry was posted in Bird evolution, dinosaur evolution, Philosophy of science, The Secret Dinobird Story and tagged , , , , , , , , , , , . Bookmark the permalink.

One Response to Feather Genes Surprisingly Early – But Misinterpreted

  1. Pingback: Denton vs. Moran on structuralism | Uncommon Descent

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s