Chapter 19: Gene Flow, Genetic Drift, and Population Structure

Welcome to Last Minute Lecture.

This free chapter overview is designed to help students review and understand key concepts.

These summaries supplement not replaced the original textbook and may not be redistributed or resold.

For complete coverage, always consult the official text.

If you want to talk about evolution, you really have to start with a fundamental truth.

Without variation, there is no change.

Exactly.

The raw material dictates everything.

I mean, if every individual were genetically identical,

natural selection would have nothing to act upon, right?

The whole engine would just, well,

seize up.

So,

our mission today is a deep dive into the very core of how that essential variation is generated, how it's structured, and how it's manipulated by those four famous forces of evolution.

We're pulling directly from the concepts that govern genetic variation, population structures, and all the mathematical modeling that underpins modern evolutionary theory.

What's really fascinating here is seeing how population genetics gives us that mathematical framework that actually supports population ecology.

We're moving beyond just looking at individual traits to analytically modeling the entire population structure, understanding not just how genes are passed down, but how they're organized within and between these breeding groups.

Okay.

Organization.

That starts with definitions, right?

When we talk about genetics acting on a group, what exactly is that group?

We define a population as basically a group of individuals that are sexually interbreeding or potentially interbreeding.

Okay.

But within any species that spread out, that large population is almost always broken down into smaller functionally distinct units we call deems.

Deems, right.

A deem is that local interbreeding group where the random mating actually happens.

So, while the whole species shares a common gene pool, it's the deems that maintain specific gene frequencies because they're exposed to very localized forces.

Oh, so it's kind of like,

if the entire species gene pool is the full pantry of ingredients available, each deem is like an individual kitchen making a slightly different local dish based on what spices are common nearby.

That kitchen analogy is perfect.

Yeah.

Yeah.

And even within those local kitchens, we find just huge amounts of available raw material.

We call this standing genetic variation and it exists in nearly all natural populations.

That sounds crucial.

It is.

It's the fuel for future adaptation.

But keeping all that variety around,

it sounds kind of exhausting metabolically speaking.

Are there downsides, consequences to keeping all that potential?

Absolutely.

And that's a key insight here.

Evolution needs variation, but actually maintaining that raw material is surprisingly costly.

We call that cost the genetic load.

Genetic load.

Okay.

That sounds pretty serious.

Does that mean like having imperfect genes?

Precisely.

It's the extent to which a population falls short of some theoretical, perfect genetic constitution.

And it's measured by the loss of individuals who don't quite hit that optimal genotype, a loss we term genetic death.

Genetic death, wow.

And, you know, genetic death doesn't necessarily mean dying young.

It could mean sterility or maybe an inability to attract a mate or just generally poor survivability compared to some theoretically perfect individual.

It's like the burden of imperfection that's necessary for future survival.

Okay.

So we can actually see this, this necessary burden, this local adaptation in that classic experiment with the yarrow plant, Achillia, right?

The one across California.

Yes, exactly.

That study really showed how adapted these deems become.

What the researchers did was they took plants from coastal areas and tried growing them at high altitudes.

They were weak, didn't do well at all.

Then they did the reverse, took high altitude plants and grew them down low.

Same result.

They performed poorly.

The genetic differences, things like height and timing of the growing season were really tightly linked.

It showed those local populations had evolved very specific, highly adaptive differences just for their own little spot, maximizing their fitness right there.

And we see this kind of population structure in humans too, right?

Based on like the traditional classifications using blood types.

We do.

Historically, human populations have been categorized into maybe five large groups, African, Caucasian, greater Asian, Amerindian, Australoid, based on things like A, B, O and R, H blood groups.

But here's the really important stat, the one everyone should remember when talking about human diversity.

Yeah.

Only about 16 % of human genetic variation comes from differences between these big continental groups.

That is the essential takeaway.

Absolutely.

Around 84 % of the genetic variation you find among humans comes from differences among individuals and populations within those same broad groups.

Wow.

84 % within.

Yeah.

So if you're looking for genetic variation, you'll find most of it right in your local neighborhood, not necessarily by comparing continents.

It really tells us that variation is fundamentally an individual and a local thing.

Okay.

So if variation is structured locally, the next natural question is how does it spread?

How does it move around?

That local adaptation must be constantly getting mixed up, right?

It is.

And that mixing mechanism is critical.

We call it gene flow.

Gene flow or gene migration.

Right.

It's simply the movement of genotypes.

That could be individuals physically moving or gemetes like pollen spreading, or even a parasite carrying alleles from one host population to another.

Okay.

The fundamental consequence is this powerful mixing of gene pools.

And crucially, this mixing acts directly to slow further differentiation between populations.

So gene flow is fundamentally working against speciation.

It slows down that evolutionary divergence that local selection might be pushing for.

Exactly right.

It's like a break.

And when you calculate the impact of gene flow on the population receiving the migrants, the math really focuses heavily on three things.

Which are?

One, the difference in the gene frequencies between the two groups.

Two, the overall migration rate, how many migrants are coming in, often called dollars.

And three, the pattern is, is it a one off event or continuous?

Okay.

The math shows that the frequency of an allele in the recipient population will eventually drift towards the frequency in the donor population.

But the speed of that change, it slows down exponentially over generations.

And it's all proportional to that migration rate, dollar dollar.

Right.

So you don't necessarily need to memorize the formula to get the main point.

The rate things get homogenized depends entirely on how much mixing there is and how big the genetic gap was to start with.

Precisely.

And we can actually see the long term continuous effect of this in studies that look at genetic exchange across large human groups, like the relationship between European Americans and African Americans.

Ah, yeah.

Using DNA markers, right.

Autosomal markers.

Exactly.

Researchers use those markers to estimate the contribution of European ancestry into about 10 different populations of African descent in the United States.

And what do they find?

The estimates showed European ancestry contributions ranging anywhere from 7 % up to maybe 23 % across those different populations.

It's a measurable consequence of continuous long -term gene flow creating these admixed populations and, well, homogenizing allele frequencies over time.

Okay.

Now let's move to the force that kind of lacks any predictable direction,

genetic drift.

Selection and flow are generally directional, pushing things one way or another, but drift is just random fluctuation.

Purely random.

And it's overwhelmingly dominant in small populations.

Think of drift as essentially statistical sampling error in evolution.

Sampling error.

Yeah.

Like if a population has,

say, 5 ,000 breeding pairs,

the sample of genes passed to the next generation will probably be pretty representative of the parent generation.

Very little random fluctuation.

Makes sense.

But if you only have two parents, that tiny sample of alleles can deviate wildly from the previous generation's frequency just by sheer chance.

It's like flipping a coin only three times getting three heads is unlikely, but not that unlikely with a small sample.

And the math confirms just how dramatic that swing can be, right?

Gosh.

The standard deviation is linked to the population size.

Yes.

It's inversely related to the square root of the number of parents, dollar rounds.

That relationship tells you that even a modest decrease in population size leads to a massive explosion in potential randomness.

So 5 ,000 parents, tiny wiggle room.

Right.

The standard deviation sigma is like 0 .005, really narrow, but with just two parents, 100 or two, the sigma jumps to 0 .25.

That's a huge potential swing in allele frequency in just one generation, almost guaranteed to be significant.

And that potential for random change, that leads directly to fixation, doesn't it?

Yeah.

If these small populations keep going, alleles can just get locked in at 100 % or disappear completely.

Exactly.

Which is why we have to talk about effective population size, usually written as dollars.

What exactly is that?

It's the number of parents who actually contribute offspring to the next generation.

And that's often far, far smaller than the total number of individuals you might count.

Ah, like if there are way more females than males.

That's a classic example.

If you have, say, 300 females, but they all breed with only three males, your census size is 303.

But the effective population size mathematically gets reduced to only about 11.

Wow.

Only 11.

Yeah.

That bottleneck of just a few breeders, in this case, the males, is what determines the strength of genetic drift, regardless of how many females there are.

It's the knowledge that matters for drift.

Okay.

This seems like the perfect place to bring in Sewell Wright.

He's one of the big three, right?

Founder of neo -Darwinian theory alongside Fisher and Haldane.

Absolutely.

And his framework, the shifting balance theory,

directly tackles this complicated interplay of drift, selection, and gene flow, especially in these small sort of fragmented populations we've been talking about.

And it revolves around his idea of the adaptive landscape.

Exactly.

Wright's whole concept is built around this adaptive landscape.

You can visualize it like a topographic map where the elevation represents fitness.

Genotypes occupy adaptive peaks.

These are the high fitness spots, and they're separated by valleys of low fitness.

Okay.

Peaks and valleys.

Now, the dilemma is that selection is pretty conservative.

It only pushes a population up the nearest peak.

It's good at climbing hills, but it's terrible at crossing valleys.

Getting from one maybe lower peak to a potentially higher one across a valley of low fitness, selection actively prevents that.

Right.

Selection keeps you stuck climbing the local hill.

So something else has to provide the freedom to move and explore.

Yeah.

That's where Drift in Small Demes becomes the hero of Wright's story.

That's precisely it.

Drift is the explorer.

Wright saw the whole process in three phases, sort of cyclical.

Lay them out for us.

Okay.

Phase one is exploration by Drift.

Genetic Drift acts on these small fragmented demes.

It randomly pushes some of them off their current adaptive peaks and maybe across those non -adaptive valleys.

It's basically exploring new genotypic territory just by chance.

Makes sense.

Random walks.

Phase two is adaptation by selection.

Once a small deme through Drift happens to land near a higher adaptive peak, then selection ticks over again.

It quickly drives that deme uphill towards the new higher fitness maximum.

So Drift finds the new hill, selection climbs it.

Exactly.

And phase three is expansion by gene flow.

The deme that successfully climbed that higher fitness peak starts to thrive and expand.

Through migration or gene flow, it eventually spreads its successful genetic combination, displacing or colonizing other demes that are stuck on lower fitness peaks.

That is such a powerful idea.

And it's quite different from, say, R .A.

Fisher's emphasis on large homogenous populations, right?

Right.

Really highlighted selection among populations, group selection as a major driver, not just selection among individuals.

Yes, that's a key distinction.

And the famous Bury experiment with Drosophila, the fruit flies, illustrates the power of Drift perfectly, even if it doesn't immediately show the adaptive part.

Tell us about that one.

They used small populations again.

Very small.

They set up 107 separate lines of flies, each starting with only 16 parents per generation.

Just eight males, eight females,

tiny dollars.

Okay.

107 lines, tiny populations.

What happened?

Did they all drift off in one direction together?

Oh, not at all.

The outcome was extreme randomness, just pure chaos driven by Drift.

By Generation 19, intense genetic drift had caused over half more than 50 of those 107 lines to reach fixation.

Fixation.

Meaning?

Meaning the allele they were tracking was either completely eliminated, frequency went to 0 % or it became the only allele left, frequency hit 100%.

And over half the populations, that specific bit of variation was just gone, lost a chance.

Wow.

Okay, that sounds entirely destructive then.

If populations are just randomly losing genetic variation to fixation, how can Wright possibly argue that Drift is advantageous?

It sounds like it just breaks things.

Ah, but here's the crucial part.

When you looked at the average allele frequency across all 107 populations combined.

Yes.

The overall average remained exactly where it started, 0 .5 or 50%.

So individual lines went crazy, but the average stayed put.

Precisely.

The key takeaway is that while Drift dramatically increases the variation between the individual populations making them very different from each other, the overall species average doesn't get pushed in any particular direction by Drift itself.

This chaotic differentiation allows the species as a whole collection of deems to explore many different parts of the adaptive landscape simultaneously.

Then group selection can potentially favor the few deems that randomly drifted onto a better peak.

Ah, okay.

So Drift creates the options and the selection among groups picks the winners.

And this emphasis on the group helps explain tricky things like altruism maybe, traits that seem bad for the individual.

Exactly.

Think about sterile insect casts like worker ants who never reproduce, or that dramatic high bouncing jump starting or pronging that antelopes sometimes do when they see a predator.

Yeah, seems risky.

As an individual, sacrificing your own reproduction or drawing attention to yourself might lower your personal selective value for sure, but a population or a deem that carries those altruistic genotypes might actually have a higher overall reproductive success rate as a group than a population that lacks them.

It favors the group even at some cost to individuals within it.

Okay, that makes sense in Wright's framework.

Yeah.

Let's shift gears slightly now.

Maybe zoom out from the mechanisms to the bigger picture, the patterns of history, phylogeography.

Yes, phylogeography.

It's essentially about reconstructing the geographical history of lineages using genetics.

So using DNA to map out where ancestors lived and moved.

Pretty much.

It examines the genetic differences among multiple populations within a species to trace their past movements and explain how they ended up distributed the way they are today.

It connects genetics to geography and history.

Cool.

Is there a good example?

A really wonderful, though complex example is the study of Asian common wheat, Triticum estivum.

Wheat.

Okay, mapping a single species like wheat across a huge complex area like Asia, especially with massive barriers like the Tibetan Plateau.

That sounds incredibly challenging.

It absolutely was.

Researchers had to figure out how wheat spread across all these diverse environments.

They proposed three main historical migration routes, largely based on historical trade routes and geography.

What were the proposed routes?

Well, one possibility was a route from Turkey towards Sichuan in China, maybe along an ancient path through Myanmar.

A second was along the famous Silk Road, connecting Central Asia and China.

And a third route likely followed the coastal areas of China and up into Korea.

Okay, plausible routes based on history.

And did the genetics actually back this up?

They did remarkably well.

They used a technique called isozyme analysis, basically looking at specific enzymes which reflect underlying genetic differences.

They analyzed something like 648 different races of wheat.

And the analysis resolved all those races into six distinct genetic clusters.

But importantly, those six clusters were ultimately derived from just three primary lineages.

Three lineages matching the three proposed routes.

Exactly.

The genetic patterns verified the migration paths suggested by geography and human history.

It's a perfect illustration of how the genetic variation within a species carries this historical signature of its past movements.

That's fascinating.

So, okay, let's try to wrap this up for our listener.

We've seen that genetic variation isn't just random noise.

It's structured in the local populations, these DEMs.

And these DEMs are constantly being shaped.

They're challenged by the homogenizing force of gene flow, mixing things up.

And also by the powerful, often counterintuitive, random effects of genetic drift, especially when populations get small.

Yeah, drift is key.

And these forces, when you put them together like Sewell Wright did, create this complex shifting balance theory where populations can explore the fitness landscape, potentially finding better adaptive peaks through a mix of chance and selection.

And finally, we can use genetics itself through phylogeography to map the historical results of all these processes.

Absolutely.

And we should always circle back to R .A.

Fisher's fundamental theorem, which really underlies so much of this.

Remind us.

The greater the genetic variation available within a population, the greater the potential rate of fitness improvement through natural selection.

Right.

Variation is the fuel.

It's the fuel, which reinforces that whole Red Queen hypothesis idea, you know, that continuous environmental change means variability itself is constantly being selected for.

Organisms have to keep adapting, keep changing just to stay in the same place relative to their competitors and environment.

The race never really ends.

So a final thought, then.

We've seen how both Sewell Wright and R .A.

Fisher gave us these mathematically solid but subtly different pictures of evolution.

Right.

Emphasizing small, fragmented deams exploring the landscape via drift.

Fisher focusing more on large, homogeneous populations adapting via selection.

Two powerful frameworks.

When you consider any species out there in the wild right now,

which model do you think is really driving its immediate evolutionary path?

Is it the Wright model or the Fisher model?

And why might the real answer be?

Well, it's probably both, maybe even depending on the day.

Something to think about.