_source

Selection#

Individual selection#

Kin selection#

Group selection#

The central dogma#

DNA -> RNA -> proteins -> life

Micro-mutations#

Gradualism#

Punctuated equilibrium#

Stephen Jay Gould and Niles Eldredge

Attacks#

"Rapid change" for paleontologists means something competely different to an evolutionary biologist#

Since paleontologists only work with fossils, they can only see the evolution of morphology#

They miss everything apart from bones/skeletons

What's the mechanism for macro-evolutionary changes?#

Macro-evolution#

Introns#

Exons#

Splicing enzymes#

Non-coding DNA#

Transcription factors#

Environmental regulation of when genes are activate
Change the context that proteins function in, when genes are expressing: if "x" is happening outside a cell, then activate this gene

Promoters#

Repressors#

Access of transcription factors to DNA#

Permanently change whether transcription factors can act
Chromatids

Examples#

Splicing enzyme mutations#

Imagine you have a gene coded for by 3 exons A B C, and messenger RNA that includes all of them + the introns.
A splicing enzyme comes along and removes the introns and exons, forming the mature protein.

What if the splicing enzyme mutates?
Say it takes all of A and half of B and makes a protein, and takes the 2nd half of B and C, and makes another protein - you get 2 completely new proteins that didn't exist beforehand - a massive change, not a protein working slightly better or slightly worse.

A gene must code for the splicing enzyme, that gene has its own promoters, exons, slicing enzyme, .., so there can be recursive regulation - 1 splicing factor can have a big consequence.

Promoter mutations#

If a promoter mutates, it'll interact with a different transcription factor.
Since multiple copies of promoters appear upstream of multiple genes, 1 promoter can mediate the expression of entire protein networks.
If a promoter mutates, it can change an entire protein network. If it changes in only some of the places, it can make a new novel network, maybe half of the original proteins.
If it mutates everywhere, maybe it interacts with a completely different transcription factor, making a new if-then clause - if x happens, make a new network of proteins.
Effects amplify - not "a protein is 1% better at binding this hormone", far bigger.

Vole vasopressin promoter mutation#

There's a promoter upstream of the vasopressin hormone - it does something with social-affiliative behaviour in males; since it's a hormone, there's a vasopressin receptor, so there's a vasopressin receptor gene, and a promoter upstram of that gene. It can come in a few different versions.

There are lots of different vole species - mountains, plains, wherever, some are monogamous, some polygamous - monogamous voles have a different promoter upstream of the vasopressin receptor gene compared to the polygamous voles.
If you change the promoter with gene therapy, you can make a polygamous vole into a monogamous one.
Change a protein -> different expression pattern (what part of the brain it winds up in etc) -> major shift in behaviour

Recent studies have shown for human males, which version of that promoter you have gives you some predictive power of how stable your social relationships will be (more/less likely to get divorced etc).

Gene that codes for dynorphin#

Dynorphin is hormone neurotransmitter that has something to do with pain perception, related to morphine.
It has a gene and a promoter - recent research has shown the more copies of that promotoer (IN RATS) predicts how readily they get addicted to some drugs

Mutation in gene for a transcription factor#

You'll get massive changes in protein networks.

If you compare the human genome to the chimp genome - what are 2% of differences?
Disproportionate share are genes that code for transcription factors - that makes sense: if you get a change in a gene for a structural protein, maybe your muscles bend a little bit one way rather than another; if a gene for a transcription factor changes, you'll have loads of new novel networks.

This was big support for King and Wilson from Berkeley, who came up with the "98%", and predicted that the most interesting changes in DNA will be in the regulatory parts, not the coding parts - eeverything since has supported that.

Genome size vs % transcription factors#

Over 100 genomes sequenced - if you line them up from short of long genomes, the more genes a species have, the higher % of genes are transcription factors:
If you have 1 gene, you only need 1 transcription factor:

1 gene    1 transcription factor
2         3 TF lets you get the max information out of them - you transcribe A, or B, or AB
3         7
4         15
5         31
n         2^n - 1

There's an exponential increase in the # of TFs you need to take advantage of all the possible combinations of networks of gene expression.
Little changes in genes that code for TFs, splicing enzymes, promoters -> big consequences

Transposons - jumping genes#

Barbara McClintock was a plant biologist, very successful, at 40 a member of the National Academy of Sciences.
One day she made a discovery that destroyed her career - molecular biology didn't exist then, all you could do to make sense of inheritance is to look at the phenotype, see if peas or people are wrinkly or not, the colours of corn kernels.
She sees a pattern of inheritance that keeps popping up - the only way for her to explain it was if genes were moving around in the DNA, so she proposed there were transposable genes - transposons, people derisively called them "jumping genes".
People laughed at her, so she disappeared to her corn field, working alone for decades, ignored.

In the 1980s, molecular techniques caught up and showed she was right.

Transposons - transposable genetic elements: genes pick up and move around.
They wouldn't have been discovered if McClintock was studying sperm whales, only in plants:
If you're an animal, you can run/crawl/fly away if things get stuff.
Plants can't move, they need a more subtle, fancier, mechanism - plant, in stress responses to a challenge, move genes around in the hopes of finding something novel/useful to get them out of trouble.
They activate transposase, an enzyme, to start transposing genes.

Animals have them too - where they were first discovered makes sense too.
You're a scientist inventing a never seen before pathogen, you inject it into people, they get sick, 2 months later they've been antibodies against it, even though it's never existed before.
How does the immune system make antibodies to fight something it's never seen before? One of the tricks is splicing genes that are relevant to making antibodies, inducable transposable events, then filtering them to see if they're good against novel pathogens.

Pathogens also do it - trypanosomes cause trypanosomiasis, and take away their surface proteins, juggle their DNA a bit, so your antibodies can't recognise them anymore.

In primates#

There's a transposable element that's predominant in primates which is most mobile when brain cells that make new neurons (neural progenitor cells), at the time they start proliferating: you get novel things neurons can do, the cells in your body that have the most to do with who you are are the least constrained by genetics - right when they're first being made, their genes are getting shuffled more than anywhere else

New if-then clauses#

Generally, a part of DNA is copied and moved randomly somewhere else, so you get big consequences.
Imagine you have an if-then clause: if dehydrated (to do with your hematocrit/how wrinkly your kidneys are/whatever, biologically speaking), then tell your kidneys to retain water.
This is ridiculous, there's no gene for this, but there are networks and ways your kidneys monitors that.
Suppose there's a transposable event, suppose the promoter moves, the "if you're dehydrated" part gets copied to upstream of the gene that makes you ovulate - that lets you do seasonal mating; 6 months of the year it's really wet, 6 months it's a drought, if your gestation period is 6 months, you want to mate in the dry season & give brith in the wet season.

Or if your gestation period is 2 months, then if you're dehydrated, don't ovulate, because your kids will be born in the dry season & starve.

This is a novel if-then clause!

Imagine you have a promoter that can tell if an individual near you smells like you, then it immediately shuts down transcription for things related to fertility - the incest taboo in loads of species.
The "smells like me" promoter can get copied to upstream of the gene that says "cooperate" - this is the start of kin selection. If there are more promoters (they smell a lot like you), then maybe you'll cooperate a lot, vice-versa - you've invented a way of taking sensory info about the degree of related-ness and turning it into whether you'll sacrifice for 1 sibling or 8 cousins.

This is again ridiculous, there aren't any promoters for "this smells like me", but there are promoters for stuff similar to it, you can imagine turning this into real biology.

Transposable genetic elements also brings up the possibility of moving around parts of genes, not just regulatory elements, but how would you do that?
Exons - for example, steriod hormones like estrogen, progesterone, testosterone, work like this:

• they enter a target cell, and bind to their receptor (lock and key - they're not made of amino acids, they have a different structure, but they still have distinctive shapes)
• each type of receptor for a type of hormone has a distinctive shape driven by it's amino acid sequence -> gene code -> etc.
• when the hormone fits into a receptor, it activates a receptor complex - on the other side of the complex there's a "confirmation" that recognises a particular promoter on the DNA, e.g. and estrogen-responsive promoter - one part of the receptor recognises the hormone (the hormone-binding domain), one part the promoter (the DNA-binding domain)
• so if you're reading the right part of a novel and secreting certain hormones and you're changing genomic effects after - environment regulating genes
• now, there's a transposable event - in steroid receptor genes the hormone-binding and DNA-binding domains are in different exons, so the estrogen receptor bit can get removed and replaced with a different hormone-binding domain
• so you have "if this other hormone is around, do this" rather than "if estrogen is around, do this"

One class of steroids are called glucocorticoids are stress hormones (hydrocortisone is the human version) - they suppress the immune system; steroidal anti-inflammatories. This is very well understood - glucocorticoids bind to the hormone- binding domain of the glucocorticoid receptor and this moves to a glucocorticoid-responsive promoter.

Imagine the progesterone receptor's hormone-binding domain gets plunked down instead - if there's progesterone around, suppress immunity. Progesterone is pro-gestational - you can suppress the immune system during pregnancy, this is good - your body doesn't have an immune reaction against it aka eat it!
An interesting consequence is once you give birth the progesterone pretty much disappears, your immune system comes back, but it's so wiped out it can come overshoot a bit and have an over-active immune system - you can get auto-immune diseases then, like a very serious form of lupus

How often are these changes good or bad?#

Sometimes genes get duplicated/tripled/etc - this area is called "copy number variantion", different people having different numbers of genes - one extra of a gene to massive duplication of sections of chromosomes. There's evidence that schizophrenia has to do with mutations in copy number variants.

In some cases, the 2nd gene is a backup - if something goes wrong in the 1st gene, the 2nd's still there doing it's job; there's a suggestion this is going on in some subsets of Alzheimers.

Or, the more copies of a gene you have, the more protein you make - recent studies show Japanese populations have more copies of a gene related to carbohydrate digestion compared to Western European populations.

Or, you can "experiment" with a copy, the other one is still doing its function - there's faster evolution in genes that are duplicated, they're more likely to stumble into some greater use without sacrificing the initial use.
Joe Thornton has done work on the evolution of genes for steroid receptors - he's shown from ansectral genes that's what happening: a lot fo what're now 2 different genes for 2 different receptors used to be duplicates of the same gene, one allowed to change, one held in place.

The "problem of irreducible complexity"#

This helps explain the "problem of irreducible complexity" - evolution can't be real because what good is half an eye?
You don't have to a rapid change from one gene to another, you can have 1 gene floating along until it happens to come up with the shape of a receptor that happens to be able to bind an earlier hormone that stumbled into existence 10k generations ago - it didn't matter that one copy was of a form with no receptor for it, since it was duplicated.
There's more and more evidence that duplicated genes describe these intermediate states - you don't have half an eye, you have the pieces in place for when something pops up that completes the picture.
Though, you can kinda have half an eye - Russell Fernald has done research on the evolution of eyes and shown there are intermediate versions of eyes, but multiple copies of genes still allow you the freedom to have looser evolving of single genes at a time - this is a critical mechanism for evolution.

Odds of good or bad mutations#

If one single base pair changes, you can have cascading changes with all of these different mechanisms (splicing enzymes, transposons etc.).

It's far more likely, with bigger changes, that mutations are bad than they're good for an organism.
This is a very stabilising mechanism for equilibrium - if one single base pair changes effects 6 proteins, or through a transcription factor, entire networks, it's not very likely you'll stumble into one change than works in all the domains, or enough of them to offset the bad effects; most of the macro-mutations are bad - there's stabilising selection against macro mutations most of the time.
This is the equilibrium in punctuated equilibrium.

The changes happen when you have circumstances that are extreme enough if doesn't matter the mutation made 10 changes and 9 of them are bit, as long as the last one will save you - it carries the other 9.
If there's suddenly an evolutionary bottleneck - very strong selection for a tiny subset of traits - it doesn't matter, if you have that trait, how much of a network you've changed, you're going to survive and no one else does.

The evolutionary record is full of examples of this, where ~1% of a population survives because of some rare trait they had, e.g. a bottleneck in selection with cheetas, because cheetas are so genetically similar to another you can graft tissue from any cheetah to another, without rejection - all cheetahs are descendent from a tiny population.

Similarly, at a couple of points in hominid evolution, there've been selective bottlenecks (like glaciers, comets, ..) where a small subset of the population, who until then had traits that weren't useful, survived.
This is circumstances of radical change.

2 critical things:

1. make a change that changes a whole network and most of the time it'll be a disaster - there'll be statis most of the time.
2. the ability to have massive macro changes in periods of selective bottlenecks

Although, this still isn't necessarily exactly right - the "rapid change" for paeleontologists is 1000s of generations for biologists.

How do you resolve punctuated equilibrium and gradualism?#

Micro-mutational changes are changes in the functions of existing proteins
Macro-mutational changes are the inventions of new proteins, new networks, new if-then clauses, serving different functions.

Human-chimp genome differences#

Back to the differences in the human-chimp genomes - looking at the micro vs. macro differences, in the immune system for example, most are micro.
Humans are far more resistant to TB, chimps to malaria - this is gradualist change, the genetic differences that explain that are micro.

For the macro differences, they're about development: one tiny difference and you'll get an organism that looks as different as human does from a chimp.
The other way around, some systems' evolution is much more driven by micro changes, some much more by macro - some of the most interesting macro stuff will be development blueprints & trajectories.

Fossil records#

There are lots of complete records of some line of species now that show punctuated equilibrium.
Though there are plenty that follow gradualist models, but the majority follow punctuated models.

Observations#

Most of the time though, it's impossible to see gradualism happening - it's hard to show if changes are incremental or rapid, but there are a few examples now of people observing evolution in organisms, ones that count as fairly rapid.

Rat carcasses in Chicago#

e.g. there are lots of rat carcasses in Chicago, from rats killed in the 1880s. Researches compared their genomes to current rats in Chicago, and showed there were a lot of changes in that time.

Darwin's finches#

People have been studying them for 50/75 years - gene distributions for traits like bill size (dictates what kind of food you can break in to), in response to environment changes, has been documented.

Diabetes resistance#

Adult onset diabetes is becoming an epidemic. Humans are built to store calories, mix that with a Westernized diet and you get obesity and diabetes.
People have studied populations that have had rapid shifts from traditional diets to Westernized ones ,like Pacific Islanders, or the iconic Pima Indians in Arizona - about half life in the USA, half Mexico. There've been far more changes to the diet in the American side, and the people in America have ~90% diabetes rates by 30 years old. The same thing has happened in Naru and Samoa, Jews who lived in Yemen and moved to Israel in the 60s/70s.

This killed off lots of people early in life, in their reproductive history, the people who were most vulnerable to diabetes. They left fewer copies of their genes. In the last 10 years or so, in some of the studied Pacific Island populations, diabetes rates are falling - there's been the selection that went into Western Europeans ~a century ago, where the people with most efficient metabolisms (that can store more calories) died.
There's been selection for worse metabolisms, within 1 century.

Domesticated silver foxes in Siberia#

They were selectively bred for tameness from wild foxes - 5/10% of them were relatively tame, only they were allowed to breed, over only 35 generations they've become as tame as dogs are.
They were bred for 1 trait only, a behavioural one - the ability to get close to them. Not just the behaviour changed - they turned into more like fox puppies, with round ears and short muzzles, and they wag their tails.
When you breed a wild animal to be tame-able, you're breeding for the ones that behave more childishly, that act more like developing animals that depend more on other individuals, for infantile traits (this is called neoteny).
Also, you can't get domestic without piebald colouration patterns.

Antibiotic resistance in bacteria#

This is very rapid change that you'll never be able to see in a fossil - bacteria don't leave fossils.

Resolution#

Both micro and macro evolutionary changes can happen at the same time - say you have evidence of a punctuated change of some trait, and single traits don't evolve at a time, they come in packages (like the foxes domestication, neoteny and coat colours).
So, there's some other punctuated change at the same time, slightly different timeframe to the first, and another one - if you have enough punctuated events happening, on a whole population level, it'll look like gradualism.

Next time#

Behavioural genetics

• someone gets adopted, do they share a trait more with their adoptive parents or their biological ones?
• non-identical twins vs identical twins
• identical twins adopted into different families at birth