- Talking about variation at single position on a nucleotide sequence among individuals

in a population
- There are 2 positions here that vary in this population sample (shown with arrows) → those
are single nucleotide polymorphisms (SNP) → very useful genetic markers
- Haplotype is like a word spelled out in the alphabet of nucleic acids, ACGT → so when we
have a unique haplotype we have a unique word
- Can see there are only 3 unique haplotypes
- Answer: C
- First thing to notice is when population is in Hardy-Weinberg → can assume the frequencies
of genotypes are in Hardy Weinburg so that in this case 4/10000 individuals is the frequency
of this genotype (little a, little a) → from that can take the square root of that and get q which
is the frequency of the mutation itself
- Then can get p (1-q) and then apply Hardy Weinberg formula for the frequencies of
heterozygotes in the population to PQ → gives 0.0392

- This is an example of a place you can apply Hardy-Weinberg to figure out what
proportion of population is a carrier
- One other application of Hardy Weinberg comes up in forensic genetics (genetics of crime
scene investigations) → DNA fingerprinting
- DNA fingerprinting often called the friend of the innocent → because if your DNA doesn’t
match the DNA left at the crime scene you’re most likely not guilty at least on basis of your
DNA fingerprint
- If DNA does match the crime scene DNA → this is where Hardy Weinberg theory is useful
- DNA fingerprinting → not looking at single locus, looking at many many loci (loci unlinked
from one another but there are many of them)
- Generally these loci tend to be SNPs or could be microsatellite alleles → generally each of
these loci has many alleles
- As long as you have lots of loci → can come up with useful way to approach this problem
- Here have this suspect in a crime that suppose there’s only 3 loci that have been
genotyped (A, B and C and each of them have 2 alleles) and blood found at the
scene matched the suspect’s DNA profile
- Probability that this match occurs simply by chance alone? → the individual may not be
guilty but happens to have the DNA profile in bold
- Need to apply the Hardy Weinberg proportions to each of the genotypes (A1, A2
genotype first)
- Should be written as 2(0.1)(0.9)
- This is frequency of B2/B2 genotype
- You’re given the frequency of the B1 allele so the frequency of B2 allele is 0.2 → 0.04 for the
frequency of the genotype itself
- C2/C2 Should be (0.7)^2

- Point is no matter what you use here → going to come up with fairly large probability of
seeing this genotype by chance alone in population when you do DNA fingerprinting → this
would most likely not be admissible evidence
- So generally when people do DNA fingerprinting for this type of thing → going to use very
low probabilities and the way to get that is to get a lot of independent locis → can go from
600 people out of 1 million to 1 out of a billion
- Just 1 form of non random mating
- When have this type of pattern occurring in a population where relatives mate with relatives
→ no longer can assume that the population is in HW proportions and have to make some
adjustments
- Says nothing about non random mating
- This is random mating
- This does not work for inbreeding
- Does not work for inbreeding because of this concept called IBD
- This pedigree has some mating between first cousins
- The colours are just tags on alleles, tags on a single locus → might end up with (1 possibility)
look at top there is dark orange allele on the maternal grand father’s side → that allele by
chance in this particular pedigree has been passed down to the next gen and etc. to the
generation at the bottom
- This is a Xerox copy of the allele that was present in the maternal grandfather
- They’re identical by descent because they’ve been passed down through this pedigree in this
way → there’s a probabilistic outcome that some of the time alleles will be identical by
descent
- We are going to be interested in the probability of identity by descent → called inbreeding
coefficient
- How to calculate probability of identity by descent (inbreeding coefficient)
- Have to identify the ancestor that could give rise to alleles that are identical by
descent
- In this example is half brother, half sister mating → only 1 individual in this pedigree who
can pass down alleles that will be identical by descent → individual A could either pass down
pink or blue allele → could pass down 1 copy each but if it passes down 2 copies of blue
allele or 2 copies of pink allele we have identity by descent → so what is that probability?
(next slide)

- Again we’re looking at the overall probability that 2 alleles are identical by descent
- First have to locate that individual that could pass down alleles that are IBD → individual A
in this case
- After that, need to diagram the path by which those are identical but could be passed
down in a way that they end up being IBD (individual I there)
- Need to look at number of individuals in that path not counting individual I → can work back
through the path so may have BAC or CAB (in this have BAC) back to I
- N=3 individuals in that path
- Apply very simple formula → the number of individuals in that path, N so have ½ raised to
the N
- The probability that they’ll be alleles that are identical by descent in this kind of pedigree is ⅛
→ this is the inbreeding coefficient for this pedigree

- Can be more complex than that → 2 complexities that can rise; 1 is individual A (who could
pass down alleles by IBD) could be inbred themselves (need to take into account the
inbreeding coefficient of that individual, so instead of having ½ raised to the N → would have
½ raised to the N times (1+inbreeding coefficient of individual A) (next slide)

- So if that individual’s inbred → have to take into account that individual’s inbreeding
coefficient
- The other possibility is there may be more than one path
- What if there’s more than 1 common ancestor? (in this case have brother sister
mating)
- Specify that individuals A and B are not inbred → their inbreeding coefficient is 0
- Hint: there are 2 paths here that you need to be concerned about → 2 individuals in that
pedigree who can pass down alleles that end up being IBD to individual I
- Answer: B
- Inbreeding coefficient in brother sister is ¼
- Here are the 2 paths; both individuals A and B could pass down alleles that end up being IBD
in individual I → inbreeding coefficient of those individuals is 0 so we can ignore FA and FB
for the moment
- We have ½ to the power of 3 and then ½ to the 3 again → add those 2 together and come up
with ¼ (⅛+⅛)
- Imagine case where had entire large population that was mating at random (no inbreeding in
past) → all of a sudden everyone in that population mates with second cousins
- Inbreeding coefficient when you have second cousin matings → probability of IBD is 1/16
- We have seen where we can calculate genotype frequencies in random mating population (on
right) but what are frequencies of the homozygotes and heterozygotes in a population with
this level of inbreeding? → pretty simple change (next slide)
- Have to tack something onto the Hardy Weinburg frequencies → for homozygotes tack on the
product F times P times Q, for heterozygotes we multiply that by (1-F)
- (Not responsible for the derivation but can look at it in the textbook)
- Should know is that in inbreeding → end up having fewer heterozygotes than expected under
Hardy Weinburg and have more homozygotes (end up with excess of homozygotes, deficit of
heterozygotes with respect to Hardy Weinburg proportions) → that’s what inbreeding does in
general
- Can imagine as raise that value of F if have severe inbreeding → going to end up with very
few heterozygotes and if inbreeding continues generation after generation → erode the
heterozygote frequencies in the population more and more and end up with more and more
homozygote
- What happens over successive generations of inbreeding is you get even further departures
from the Hardy Weinburg → one reason why it matters is because of inbreeding depression
(next slide)

- Are there any change in allele frequencies under inbreeding? → no it does not, it causes a
change in genotype frequencies but can see the allele frequencies are exactly the same under
Hardy Weinburg
- Just the genotype frequencies that are changing under inbreeding
- There are quite a few recessive deleterious alleles in the population and in many
cases can cause diseases like listed on slide
- Point is if you raise the level of inbreeding in a population and consequently the level of
homozygosity in a population → going to end up with more and more of these diseases being
expressed in a population
- Not the only thing → many diseases are single locus diseases but a lot of fitness is determined
by multiple loci and if have mildly deleterious recessive mutations that build up and their
joint effects is to reduce vigor or viability and reproductive success in some way → going to
get lots of decline in health when have lots of inbreeding in a population or decline in fitness
of the average individual in a population
- Across Y axis is infant mortality rate of the 19th century
- Can see on average as you look at infant mortality of individuals that are more and more
inbred → have an increase in infant mortality
- Factors that can actually change allele frequencies in populations
- Next lectures will discuss natural selection, just discuss these 3 right now
- Remember these things can interact with one another, but we will just look at them
independently at first
- CF is disease that arises from mutation → that mutation is circulating in population already
(not usually arising each generation) → this is because mutation rates especially at base pair
level
- Several forms of this mutant allele but if consider a single base pair mutating and causing CF
in homozygote → mutation rate is low
What is the mutation rate at the level of the base pair per individual per generation?
- Answer: D, 1 out of a billion
- We have roughly 3 billion nucleotides in our genome
What about per gene? Would these numbers apply?
- Say average gene has 1000 base pairs → then going to have a lower number, maybe 1 out of
10 million per gene (gene will mutate 1 out of 10 million times)
- When you have a small number (say here 10 to minus 5) per generation → if start out with
particular allele frequency at time TPt, the ones that don’t mutate are 1 minus that mutation
rate
- If you’re following PT over time → going to have progressively smaller number and going to
look like that
- But if look at y axis (100,000 generations) → mutation rate by itself does not cause a very
large change in allele frequencies → basically if we’re making short term forecast about how
a population will evolve, may not need to consider the mutation rate
- Natural selection especially work with mutation to change population’s genetic
composition quite quickly, but mutation alone has relatively slow effect
- Migration is a bit like mutation in sense that it introduces new alleles into the gene
pool from somewhere else
- When think about mutation → intro is de novo
- When think about migration → those new alleles are coming into the population from another
population
- Need to specify in order to make general statements about the effect of migration in a
population → is it bidirectional? Bidirectional?
- If unidirectional → it’s a bit like mutation (getting intro of alleles every generation into a
population that weren’t there before)
- Migration rates in contrast to mutation can be fairly high
- If have lots of genetic exchange back and forth between individual populations → that’s
homogenizing force (keeps populations from diverging from one another)
- This is unidirectional

- Bidirectional
- Usually multiple populations in a species and when talking about homogenizing → would
need some complex pattern with lots of arrows correcting individual subpopulations
- When have genetic drift → usually talking about small populations
- This is extreme example of only 10 individual rabbits with diff genotypes
- Just assuming nothing else going on but small population size, no mutation, no migration, no
natural selection → just genetic drift
- Say held a population at this size for 3 generations like this and simply had rabbits reproduce
and end up choosing 10 progeny to do that again → would not be surprised to find the allele
frequencies, B gene starts out equal and might change by chance alone in the next generation
→ and again 3rd population you might not be surprised to lose that little B allele
 - This is simple diagram of genetic drift → population just drifting
- Small population size that brings us genetic drift into play

Why does that matter? Consider that again in a diagram except this time 6 diff populations
will follow through time
- N=10 at the top, N=500 at bottom
Focus on top
- Graphical representation of the thing we just looked at → see drift there through time →
overall outcome is the divergence of these populations from one another → they diverse due
to chance and the loss of genetic diversity
What if had migration occurring in that top diagram? Would these populations diverge as
quickly from one another?
- If had very high migration rates between populations → these populations would diverge less
quickly
- Migration would slow down the loss of alleles because there would be effectively a
larger population
Focus on bottom
- Population larger, Don’t see the same extreme genetic drift