Welcome back, everybody. Today is the fu...

by Unattributed

Audio version created with Paper2Audio.

Listen on Paper2Audio

Audio by Paper2Audio

Welcome back, everybody. Today is the full Exam Two review. We're going down your study guide one item at a time: population genetics, selection, quantitative traits and heritability, evolution and speciation, phylogenetic trees, non-Mendelian inheritance, dosage compensation, and developmental genetics.

I'll walk through the concepts, work the calculations out loud, and run your lecture practice questions along the way. If you're driving or walking, just listen. If you're at a desk, grab a pencil for the math sections, because the math is where people lose points.

Let's start with the biggest calculation block.

Part One: Allele and Genotype Frequencies

A population is a group of individuals that interbreed in the same area. Population genetics is the study of variation in that group. About ninety percent of variation comes from S.N.P's, single nucleotide polymorphisms, and most variation is silent, meaning it has no phenotype.

Two vocabulary words. A site is polymorphic if the population has two or more alleles at it. It's monomorphic if there's only one. The gene pool is all the alleles in the population.

Now the two formulas. Allele frequency is the number of copies of that allele divided by the total number of alleles. Genotype frequency is the number of individuals with that genotype divided by the total number of individuals.

The trap is that every individual carries two alleles, so every individual counts twice when you're counting alleles.

Here's the frog example from lecture. Sixty-four dark green frogs, genotype big G big G. Thirty-two medium frogs, big G little g. Four light frogs, little g little g. A hundred frogs total.

Genotype frequency of light frogs is four out of a hundred, so four percent.

Allele frequency of little g: each light frog has two copies, so four times two is eight. Each medium frog has one, so that's thirty-two. Eight plus thirty-two is forty. The total number of alleles is two hundred, a hundred frogs times two. Forty over two hundred is twenty percent.

Now the practice question about alligator gar in Buffalo Bayou. Ten monsters, forty average, thirty small, with an incompletely dominant G allele. Incomplete dominance means all three genotypes are distinguishable, so we can count directly. Monsters are G G, average are G g, small are g g. That's eighty fish and a hundred sixty alleles. G copies: ten times two is twenty, plus forty, which is sixty. Sixty over a hundred sixty is zero point three seven five, so about zero point three eight. Answer B.

Part Two: Hardy-Weinberg Equilibrium

Hardy-Weinberg equilibrium means allele and genotype frequencies don't change from generation to generation. It's the baseline expectation. We use it to ask whether something, like selection or migration, is acting on a population.

There are five assumptions, and you need to know them cold. One, no new mutations. Two, no genetic drift.

Three, no migration. Four, no natural selection. Five, random mating. Break any one and you've violated equilibrium.

Your slide summary lists the violations as natural selection, genetic drift, migration, non-random mating, and new mutations.

The equation is p squared plus two p q plus q squared equals one, and p plus q equals one. P is the frequency of the big G allele, q is the frequency of little g. P squared is the frequency of G G, q squared is g g, and two p q is the heterozygotes.

Lecture example: p is zero point eight and q is zero point two. G G is zero point eight squared, sixty-four percent. g g is zero point two squared, four percent. Heterozygotes are two times point eight times point two, thirty-two percent.

Those match the frog numbers. And without selection or migration, those ratios hold forever.

Now cystic fibrosis. It's recessive and affects about one in twenty-five hundred babies. The affected people are the homozygous recessives, so q squared equals one over twenty-five hundred, which is zero point zero zero zero four.

Take the square root. Q is zero point zero two, so two percent. That's answer A.

Follow-up: what percent are carriers? P is zero point nine eight. Two p q is two times point nine eight times point oh two, about zero point zero three nine, roughly four percent. Answer C.

The pattern for any Hardy-Weinberg problem is to start with the thing you can see. Usually that's the recessive phenotype, which gives q squared. Take the square root for q, subtract from one for p, and then get everything else.

Part Three: Selection and Fitness

Selection means differences in reproductive success. Higher survival, better mating chances, or higher fertility raise an allele's frequency over time, and that changes how common phenotypes are.

There are four types, and they show up as curves, so know how each reshapes a distribution.

Directional selection favors one extreme phenotype, and the whole curve shifts. The example is moth wings during the industrial revolution, light wings shifting to dark.

Balancing selection maintains two or more alleles because of opposing selective forces. The classic case is heterozygote advantage, like sickle cell.

Disruptive, or diversifying, selection favors two or more phenotypes. The curve splits into two humps. It's common in diverse environments with different fitness, like snail shells, and it increases genetic diversity.

Stabilizing selection works against the extremes, so the curve narrows around the middle. The example is egg number, and it decreases genetic diversity.

Memory trick on diversity: disruptive increases it, stabilizing decreases it.

Now fitness, written as w. Fitness is the likelihood that a genotype reproduces. Relative fitness is average reproductive success scaled so the best genotype equals one. In lecture, A A had five offspring, A a had four, and a a had one.

A A is five over five, which is one. A a is four over five, point eight. And a a is one over five, point two.

Mean population fitness, w bar, is p squared times w A A, plus two p q times w A a, plus q squared times w a a. With both alleles at point five: point twenty-five times one, plus point five times point eight, plus point twenty-five times point two. That's point twenty-five plus point four plus point oh five, which is point seven.

After one generation of selection, each genotype's new frequency is its starting frequency times its fitness, divided by w bar. A A is point twenty-five times one over point seven, about point three six. A a is point five times point eight over point seven, about point five seven. And a a is point twenty-five times point two over point seven, about point oh seven.

For the allele frequency, take A A plus half of A a. That's point three six plus about point two eight five, which is about point six four. So the A allele went from point five to point six four in a single generation.

Part Four: Drift, Migration, and Mating Patterns

Genetic drift is random change in allele frequency, and it's most common in small populations. Two flavors you need.

A bottleneck is a population collapse, like a natural disaster. Survivors are random, not selected, so alleles get lost by chance. Genetic diversity drops, and the population might even become monomorphic. In the lecture picture, blue marbles randomly survive and the rest vanish.

The founder effect is when a small subpopulation migrates and starts a new colony. The new population only carries the founders' alleles, so diversity is lower. The example is the Amish. Roughly three couples founded a population of about eight thousand, and Ellis-van Creveld syndrome shows up in about seven percent of them.

Migration means gene flow, the transfer of alleles between populations. It decreases diversity between populations but increases diversity within populations.

Non-random mating has two pieces. Assortative mating is positive when individuals prefer similar phenotypes and negative when they prefer dissimilar ones. Then there's outbreeding, mating between unrelated individuals, versus inbreeding, mating between genetically related individuals. Inbreeding increases the number of rare homozygous alleles, and since many diseases are recessive, that leads to inbreeding depression, reduced fitness.

Quick check. Does selection change allele frequency randomly or directionally? Directionally. Drift is the random one.

Part Five: Mutation and the Source of Variation

Evolution needs three things. Traits vary in a population, traits are heritable, and some traits increase fitness. Beneficial traits then rise in frequency, and if populations stay separated long enough, you get new species.

Mutation is the ultimate source of variation. Most mutations are neutral, with no phenotype or fitness effect. They accumulate at a steady rate, mostly due to genetic drift, which is why most mutations are not adaptive. That steady rate is also what makes molecular clocks possible, and we'll come back to that. Mutations are caused by D.N.A replication errors, radiation like U.V, and viral infections. Bad mutations get removed from the population by purifying selection.

Three ways variation becomes new biology. A point mutation, like an A to G change, can modify a protein's activity. A duplication gives you two copies of a gene, one keeps the original function and the other is free to evolve a new function. And expression changes: a transposable element or S.N.P lands near a gene and turns it on in a new tissue.

Conservation is a clue to function. When you align protein sequences from related species, the invariant positions are highly conserved. Mutations there break the gene and get removed by purifying selection, so conserved sites predict functionally essential regions, like the enzyme active site in the folded protein.

Now the duplication vocabulary, which is explicitly on your study guide. Take the ancestral mammal globin gene. It duplicated into alpha globin and beta globin.

Then horse and human split. Alpha and beta within one species are paralogs. Duplication within a genome creates paralogs.

Human alpha and horse alpha are orthologs, the same gene in different species. Paralogs equals within species, orthologs equals between species.

Part Six: Speciation

Three geographic modes. Allopatric speciation: a barrier completely separates two populations. Parapatric: a partial barrier, with a hybrid zone in the middle. Sympatric: the same area, with no geographic separation at all.

When do two populations become two species? Still debated, and most definitions have exceptions. The big idea is reproductive isolation, where different species can't or won't mate.

Prezygotic isolation means no breeding happens at all, like different mating calls. Postzygotic isolation means fertilization happens but offspring aren't viable or fertile, like genomic incompatibilities or sterile offspring. Postzygotic problems are often related to gene copy number differences, incompatible S.N.P's or transposable elements, and chromosome differences. Remember inversions cause meiosis defects and semi-sterility.

Examples. Meadowlarks: western and eastern look almost identical, have mostly different ranges, and don't interbreed where they overlap because of different mating calls. That's prezygotic isolation. African cichlids: around ten thousand species in the Great Lakes, but they breed in captivity and produce viable offspring, with lots of evidence of past and present hybridization.

So the definition has exceptions. And the poison dart frog slide shows that morphology alone can be a poor definition.

Part Seven: Phylogenetic Trees

A phylogeny is a reconstruction of past relationship history for a group of species. Anatomy of a tree: tips or leaves are the species, branches connect them, and nodes represent common ancestors. Data come from D.N.A or protein sequences and from traits like morphology, and close relatives have more similarity.

Reading a tree: relatedness means how recently two groups share a common ancestor, not how close they sit on the page. In the lecture mammal tree, the dolphin question asks which species is most closely related to dolphin. Dolphin's closest relative on the tree is whale, so among the options, blue whale is the answer, D.

Building trees uses parsimony: the best tree has the fewest mutations. Take each possible topology and count the mutation events needed to explain the aligned sequences. In the lecture example, three possible trees needed four, six, and five mutations. The four-mutation tree wins.

In the practice question with species A through D, the informative sites favor grouping A with B and C with D. Counting steps, Tree 1 needs seven changes, Tree 3 needs eight, and Tree 2 needs nine. Tree 1 is most parsimonious, answer A.

Tree search gets huge fast. Three species have three possible topologies. Four have fifteen.

Six have nine hundred forty-five. Ten have about thirty-four million. Twenty have about eight point two times ten to the twenty-first. So you need software to search, and the best tree is not always found.

Molecular clocks: they're based on mutation rate, new mutations per nucleotide per generation. Because mutation rate is mostly constant, the number of mutations between species tells you how many years since the last common ancestor. Time calibration places nodes at specific time points, guided by the fossil record.

Part Eight: Quantitative Traits, Q.T.L, and gwas

Complex traits are determined by multiple genes and/or environmental factors. The multiple-gene part is called polygenic. They usually show a gradient of phenotypes that can be measured, like height, speed, and weight. Those are quantitative traits.

Human height shows the scale. A large study of about five point four million individuals found twelve thousand one hundred eleven independent S.N.P's associated with height, clustered in seven thousand two hundred nine genomic segments covering about twenty-one percent of the genome.

Two ways to find the genes. Q.T.L mapping, quantitative trait locus mapping, uses controlled crosses. Cross a big-fruit parent with a small-fruit parent, each carrying lots of unique S.N.P's. The F.1 is a mix, recombination shuffles the chromosomes over many crosses, and then you compare the sequences of the biggest offspring against all or the smallest.

The region with the largest allele frequency difference is the locus. The catch: it may contain many genes, so which one matters?

gwas, genome-wide association study, is used where controlled crosses aren't possible, like humans. It tests the statistical association between every S.N.P and a trait, and requires thousands to hundreds of thousands of individuals. For each S.N.P you plot trait value against genotype and fit a regression. A steeper slope means a larger S.N.P effect size.

Multiple testing: if you test a hundred thousand S.N.P's at p equals point oh five, about five thousand will be false positives. So you correct. The Bonferroni cutoff is point oh five divided by the number of tests, here point oh five over a hundred thousand, which is five times ten to the minus seventh. Very conservative.

Results display as a Manhattan plot, with S.N.P's along the chromosomes and peaks above the cutoff line. Peaks cluster because of linkage. The S.N.P itself may not cause the phenotype, it's just near something that does. Regions may contain many genes, and it's not clear which matter.

Caveats from lecture: gwas often finds very small effect sizes, some studies are questionable, and historically they've had a Eurocentric bias.

Part Nine: Heritability

Total phenotypic variance equals genetic variance plus environmental variance, V P equals V G plus V E. Heritability is the fraction of variation due to genetic variation, V G over V P.

Lecture example: inbred mice are mostly homozygous, so V G is zero and their variance is all environmental, fifteen grams squared. Wild mice have a total of twenty-two. So V G is twenty-two minus fifteen, seven grams squared. Heritability is seven over twenty-two, about point three two.

Practice question: inbred strain variance point five, heterogeneous strain point seven. Genetic variance is point seven minus point five, point two. Answer A.

Narrow-sense heritability uses correlations between relatives. It equals the observed correlation divided by the expected correlation. Expected values: identical twins one point zero, siblings point five, parent and offspring point five, uncle and niece point two five. Chicken example: egg weight correlation between mothers and daughters is point three, expected is point five, so point three over point five is point six.

Big conceptual trap, and it's a practice question. Height heritability of point eight does not mean eighty percent of your height is due to genetics. It means eighty percent of the variation in height in that population is due to genetic variation. Answer C. Heritability describes populations, not individuals.

Gene-by-environment interaction: the effect of a genotype depends on the environment. In the soil example, H H plants are tall in both soils, H h are medium in standard soil but tall in supplemented soil, and h h are short in standard soil but tall in supplemented soil. The environment has a big effect on the genotypes that are inefficient, and little on the ones already at maximum.

Part Ten: Non-Mendelian Inheritance

Mendelian inheritance assumes genes act through the offspring's own expression, pass unaltered, obey segregation, and obey independent assortment. Breaking each gives a non-Mendelian pattern: maternal effect, epigenetic inheritance, cytoplasmic inheritance, and linkage.

Maternal effect: the mother's genotype alone determines offspring phenotype, because gene products accumulate in the early embryo. The embryo doesn't make its own gene products for a while and relies on mom's. Nurse cells provide proteins and nutrients to the egg, so variants in mom's genome affect the egg. The egg's own genotype doesn't matter.

Classic example, snail shell coiling. Dextral is right-handed and sinistral is left-handed, determined by mom's genotype. Dextral D is dominant. A d d mother makes sinistral offspring whatever the father is. Practice question: sinistral female d d crossed with dextral male D D. Offspring are all genotype D d, but all sinistral, because the phenotype follows the mother's genotype.

Answer A. In the F.2, a D d mother gives all dextral offspring in a one to two to one genotype ratio, and the genotypes are irrelevant. The mechanism is that D and d control mitotic spindle orientation in early cleavage.

In flies, the body axis is a maternal effect. Nurse cells deposit bicoid on the anterior side and nanos on the posterior. A female with no bicoid has a hundred percent of offspring with no heads, two tails.

Epigenetic inheritance: chemical modification to D.N.A, like methylation, can alter gene expression, sometimes can be inherited, and is reversible. Methylation at a promoter turns a gene inactive. Transgenerational epigenetic inheritance in humans is a controversial frontier, and the Dutch famine studies are an example, with evidence that's circumstantial at best.

X-inactivation. In placental mammals, females have two X chromosomes and one is randomly inactivated, condensed into a Barr body. It happens randomly and mid-development, producing patchwork phenotypes like calico cats. A heterozygous cat has a black allele on one X and an orange allele on the other, and different patches inactivate different Xs.

Cells can count their X chromosomes, keeping only one active. Barr bodies: normal female X X has one, normal male X Y has zero, Turner X O has zero, triple X has two, and Klinefelter X X Y has one. It's number of X chromosomes minus one.

Calico practice question: calico female X B X b crossed with orange male X b Y. Daughters get either X B or X b from mom plus X b from dad: calico and orange. Sons get X from mom and Y from dad: black or orange. So one orange female, one calico female, one black male, one orange male. Answer A.

Dosage compensation across animals, listed on your guide. Placental mammals: one X is randomly inactivated in females. Drosophila: the single X in males has two times expression. C. elegans: X X individuals reduce expression fifty percent from both X chromosomes. One flag: your slide labels the C. elegans sexes oddly, but the mechanism you need is the fifty percent reduction on both Xs in the X X sex.

Cytoplasmic, or extranuclear, inheritance: genes in mitochondria and chloroplasts. The mitochondrial genome is circular, about seventeen thousand base pairs, with thirteen genes plus T.R.N.A's and R.R.N.A's. Chloroplast genomes are about ten times bigger, tobacco is about a hundred fifty-six thousand base pairs, with roughly one hundred ten to one hundred twenty genes. An egg has around a hundred thousand times more mitochondria than sperm, so in animals all mitochondria come from mother.

Heteroplasmy is a cell having more than one organelle variant. In plants, a variegated leaf has green and white chloroplasts, which sort randomly into daughter cells. The crosses to memorize: white female by green male gives all white.

Variegated female by green male gives green, white, and variegated. Green female by white male gives all green. Green female by variegated male gives all green. The rule: the offspring follow the mother.

Human mitochondrial diseases: over two hundred known, often involving A.T.P-rich tissues like neurons and muscle, with a maternal inheritance pattern. Heteroplasmy gives a wide phenotype spectrum within families. Examples: Leber hereditary optic neuropathy involves respiratory chain genes, neurogenic muscle weakness involves the Atpase six gene, and mitochondrial myopathy and maternal myopathy with cardiomyopathy both involve a T.R.N.A for leucine.

Distinguishing maternal effect from cytoplasmic inheritance: maternal effect means mom's nuclear genotype supplies products to the egg. Cytoplasmic inheritance means the genes themselves are in organelles and passed through the egg.

Part Eleven: Developmental Genetics

Four cell activities in development: division makes more cells, migration moves cells around, death is selective cell loss, like removing webbing between fingers, and differentiation makes specialized cells.

Differentiation is driven by progressive activation of cell-type-specific genes, directed by transcription factors that turn on specific sets of genes. Think of the hematopoietic stem cell branching into lymphoid and myeloid lineages. Differentiation is usually permanent, but transcription factors can return cells to stemness in the lab as induced pluripotent stem cells, iPS cells. The example is a skin biopsy turned into neurons or muscle to study disease. The protocol adds transcription factors and signaling proteins in steps to recreate development in a dish. And in C. elegans, the fate of every one of the nine hundred fifty-nine somatic cells is known.

Four phases of development: formation of body axes, segmentation of the body, determination of structures within segments, and cell differentiation.

Segmentation: the embryo forms distinct grooves, and each segment has a unique identity and cell fates. Regional identity comes from morphogen gradients and homeotic genes. Morphogens are proteins that regulate developmental fate in a concentration-dependent manner, forming gradients between secreting and receiving cells.

Overlapping gradients give each cell a different combination, producing stripe and spot patterns. Practice question: molecules that convey positional information are morphogens, answer A.

Homeotic genes establish regional fate. They're transcription factors with a homeobox domain, and they're co-linear: arranged on the chromosome in the same order they're expressed along the body axis, with expression maintained in adults. In flies there are two complexes, Antennapedia and bithorax. Misexpression causes homeotic transformation. Antennapedia mutants grow legs where antennae should be, and bithorax mutants turn the haltere into a wing. In vertebrates they're Hox genes, four clusters in mice, A, B, C, and D.

Practice question: a novel Hox gene, highly homologous to Hox thirteen. Because of co-linearity, Hox thirteen is expressed at the posterior end, so the gene is near the end of the spine closest to the tail. Answer B.

Experimental approaches, which are explicitly on your guide. Gene discovery has two directions. Forward genetics goes phenotype to gene: introduce random mutations, like with E.N.U, screen progeny for a phenotype of interest, then map the mutation to identify the gene. Reverse genetics goes gene to phenotype: pick a gene, disrupt it with R.N.A.i, morpholinos, or crisper Cas nine, and look for a phenotype.

Gene localization: in situ hybridization uses labeled R.N.A probes to find M.R.N.A in cells. Immunostaining uses labeled antibodies to detect proteins. And a transgene fuses G.F.P to a protein and adds it to the genome with genome editing, so you can watch the protein glow.

Quick memory aid: forward starts with the phenotype, reverse starts with the gene. R.N.A probe means M.R.N.A, antibody means protein.

Closing

Here's your final checklist. Can you calculate allele and genotype frequencies, with and without Hardy-Weinberg? Can you list the five assumptions and the violations? Do you know the four selection types and how to calculate relative fitness?

Can you find heritability from variance and from correlations, and explain what it means? Paralogs versus orthologs. The speciation modes and the pre- and postzygotic barriers. Tree anatomy and the most parsimonious tree.

Each non-Mendelian pattern, with a cross you can predict. Dosage compensation in mammals, flies, and worms. And the developmental terms and experimental methods.

If any of those made you hesitate, go back to that section and redo the practice problem on paper. Good luck on the exam. You've got this