1. The Structure, Function, and Evolution of Biological Systems Instructor: Van Savage Spring 2010 Quarter 4/27/2010 Instructor: Van Savage Spring 2010 Quarter 4/27/2010

2. Outline 1.Finite size populations and genetic drift 2.Coalescence 3.Understanding directional and nondirectional forces 3. General Diffusion Equation 4. Biology a. Bacterium’s use of 3 (physical constraint) b. Population genetics--combining selection and drift in evolution (Conceptual analogy) Derive equations in this context 1.Finite size populations and genetic drift 2.Coalescence 3.Understanding directional and nondirectional forces 3. General Diffusion Equation 4. Biology a. Bacterium’s use of 3 (physical constraint) b. Population genetics--combining selection and drift in evolution (Conceptual analogy) Derive equations in this context

3. "But as much as we would like to take a unified view of nature, we keep encountering a stubborn duality in the role of intelligent life in the universe, as both subject and student. We see this even at the deepest level of modern physics." Steven Weinberg

4. Genetic Drift--Nondirectional force Drift--by random, transient, non-genetic events, individuals that would be highly reproductive in a “repeated” experiment, are lost to death. Imagine a population of identical individuals that are being chosen for mating. Random chance that a given individual will be chosen. Drift--by random, transient, non-genetic events, individuals that would be highly reproductive in a “repeated” experiment, are lost to death. Imagine a population of identical individuals that are being chosen for mating. Random chance that a given individual will be chosen.

5. Genetic Drift Model this by randomly sampling from entire population (Wright-Fisher model). Population size, N, is constant, and individuals are randomly selected for mating Each individual has 1/N chance of reproducing. We get a binomial tree that depends on frequency, p, and total population size, N. Model this by randomly sampling from entire population (Wright-Fisher model). Population size, N, is constant, and individuals are randomly selected for mating Each individual has 1/N chance of reproducing. We get a binomial tree that depends on frequency, p, and total population size, N. Generation 0

6. Genetic Drift time frequency, p 1 So, rate of spread of the width of distribution is~p(1-p)/2N

7. Coalescence: Look backwards in time

8. Quick results 1.P(fixation)=p 2.P(fixation new individual mutant)=1/2N 3.Total fixation rate of mutations=2Nμ(1/2N)=μ 4.Probability of coalescing t+1 generations back is (1/2N)e -t/2N 1.P(fixation)=p 2.P(fixation new individual mutant)=1/2N 3.Total fixation rate of mutations=2Nμ(1/2N)=μ 4.Probability of coalescing t+1 generations back is (1/2N)e -t/2N

9. Using last result 1. =2N 2.StDev(T)=2N 1. =2N 2.StDev(T)=2N

10. General diffusion equation and combining selection and drift

11. Directional Forces: Go with the flow

12. Directional Forces 1.Crowd running towards a celebrity or away from a fire. 2.Pushing or rolling any ball or object 3.A river flowing towards the sea or ocean 1.Crowd running towards a celebrity or away from a fire. 2.Pushing or rolling any ball or object 3.A river flowing towards the sea or ocean Position t0t1t2t3t4

13. Nondirectional forces: No flow

15. Nondirectional forces v 1.Lost in a crowded intersection 2.Drop of dye in water 3.Smoke 4.Choosing each step after flipping a coin

16. Net Flow--Directional Forces f(t,i) is abundance of or probability of being in bin i at time, t. v(i) is speed of flow out of bin i. f(t,i) is abundance of or probability of being in bin i at time, t. v(i) is speed of flow out of bin i. i-1 i i+1 Net Flow=Flow In-Flow Out =Flow from left(i-1->i) - Flow to right(i->i+1) ->f(t+1,i)-f(t,i)=v(i-1)f(i-1,t)-v(i)f(i,t) Continuum limit: ->df/dt=-d(vf)/dx (i.e., distance=velocity*time) x f equilibrium boundary(wall)

17. Net Flow--Nondirectional Forces i-1 i i+1 Net Flow=Flow In-Flow Out =Flow from left (i-1->i)+Flow from right (i+1->i) -Flow to right (i->i-1)-Flow to left (i->i+1) ->f(t+1,i)-f(t,i)=D(i-1)f(i-1,t)+D(i+1)f(i,t)-2*D(i)f(i,t) D(i) is the diffusion rate Continuum limit: ->df/dt=d 2 (Df)/dx 2 (Second derivative) Local process and affects width of distribution, not mean x f Equilibrium (D=constant)

18. Global signature of diffusion Random walk x(t+1)=x(t)±1 ->x 2 (t+1)=x 2 (t)+2x(t)+1(1/2 of time) =x 2 (t)-2x(t)+1(1/2 of time) =(1/2)* +(1/2)* = +1= +2 Iterating this gives: =Number of time steps~t Random walk x(t+1)=x(t)±1 ->x 2 (t+1)=x 2 (t)+2x(t)+1(1/2 of time) =x 2 (t)-2x(t)+1(1/2 of time) =(1/2)* +(1/2)* = +1= +2 Iterating this gives: =Number of time steps~t

19. Diffusion properties and simulation http://web.mit.edu/course/3/3.091/www/diffusion/ Internal dynamics of diffusion are not immediately obvious from global behaviors, whereas it is for directional forces Nondirectional force (diffusion) affects width of distribution, and directional forces affect the mean. Diffusion process is NOT time reversible. Initial conditions are forgotten. http://web.mit.edu/course/3/3.091/www/diffusion/ Internal dynamics of diffusion are not immediately obvious from global behaviors, whereas it is for directional forces Nondirectional force (diffusion) affects width of distribution, and directional forces affect the mean. Diffusion process is NOT time reversible. Initial conditions are forgotten.

20. “Central to our feelings of awareness is the sensation of the progression of time. We seem to be moving ever forward, from a definite past to an uncertain future. More often, we feel ourselves to be helpless spectators - perhaps thankfully relieved of responsibility, - as, inexorably, the scope of a determined past edges its way into an uncertain future. Yet physics, as we know it tells a different story. All the successful equations of physics are symmetrical in time. They can be used equally well in one direction of time as in another." Roger Penrose

21. 2. Combined Effects Person trying to walk north (directional) through a busy intersection (nondirectional) Net Flow=Directional Flow+Nondirectional Flow Person trying to walk north (directional) through a busy intersection (nondirectional) Net Flow=Directional Flow+Nondirectional Flow Diffusion Equation (Also known as Kolmogorov forward equation)

22. Often, v and D are constant, so:

23. 3. Physics--Brownian Motion Molecule in glass of water is analogous to our person walking through a crowd. Since molecule is so small (mass is so little), gravity’s effect (directional force) is negligible. Hence, Molecule in glass of water is analogous to our person walking through a crowd. Since molecule is so small (mass is so little), gravity’s effect (directional force) is negligible. Hence, (Heat Equation)

24. The principle of generating small amounts of finite improbability by simply hooking the logic circuits of a Bambleweeny 57 Sub-Meson Brain to an atomic vector plotter suspended in a strong Brownian Motion producer (say a nice hot cup of tea) were of course well understood - and such generators were often used to break the ice at parties by making all the molecules in the hostess's undergarments leap simultaneously one foot to the left, in accordance with the Theory of Indeterminacy.Many respectable physicists said that they weren't going to stand for this - partly because it was a debasement of science, but mostly because they didn't get invited to those sort of parties.--Douglass Adams

25. 4. Biology—Magnetotactic Bacteria

26. Magnetotactic Bacteria Weigh so little that gravity is negligible, and they do not know which way is down. Better conditions at bottom (oxygen pressure) They have internalized enough magnetite particles so that earth’s magnetic field can just overcome nondirectional forces of Brownian Motion. Since magnetic fields go into earth, they can now sense down. So, they “solve” problem from previous slide! Bacteria in north and south are polarized differently.

27. Apply magnetic field to diffusing particles This gives a directional force, and since magnetic force is much stronger than gravity, this is not negligible. Must return to full equation. v depends on strength of magnetic field

28. Types of multidisciplinary influences Physical Process Conceptual and Mathematical Analogy Magnetotactic bacteria Evolution and Population Genetics -> Combine natural selection and genetic drift

29. 4. B. Combining Selection and Drift We can also understand process of evolution by means of diffusion equation. Requires different sort of extension to biology. It’s not just understanding how biological organisms use and are constrained by physics, but it’s using analogies to mathematical physics to understand biological problems.

30. Selection--Directional Force Let a population (wild type) suddenly have a few individuals with a mutation that forms a new allele. If fitness (as measured by growth rate--number of offspring per individual per generation that survive to next generation) of wild type is normalized to 1, and mutants have fitness 1+s Let a population (wild type) suddenly have a few individuals with a mutation that forms a new allele. If fitness (as measured by growth rate--number of offspring per individual per generation that survive to next generation) of wild type is normalized to 1, and mutants have fitness 1+s Population Size time Wild type Mutant

31. Position space, x, is replaced by frequency space, p, for frequency of mutants. Velocity of selection force is~ p(1-p)s time, t frequency of mutants, p 1 If population size is fixed (finite resources), only a matter of time, until mutant takes over (fixation).

32. Genetic Drift time frequency, p 1 So, rate of spread of the width of distribution is~p(1-p)/N

33. Strong Analogy 1. Selection Gravity Pushed by marathon 2. Drift Brownian motion Crowd at intersection 3. Small organisms Small populations (Brownian>>Gravity) (Drift>>Selection) 4. Large organisms Large populations (Gravity>>Brownian) (Selection>>Drift) 1. Selection Gravity Pushed by marathon 2. Drift Brownian motion Crowd at intersection 3. Small organisms Small populations (Brownian>>Gravity) (Drift>>Selection) 4. Large organisms Large populations (Gravity>>Brownian) (Selection>>Drift)

34. Equation for Population Genetics Questions we can answer using this equation. 1. Is mutant population likely to go extinct or take over population (fixation)? 2. How long does it take before extinction or fixation occurs? 3. For a given N and s, how large does p 0 need to be before mutants are likely to take over. p 0 is initial frequency of mutants in the population.

35. More proper derivation Taylor expand in p around epsilon to get Kolmogorov forward equations probability density of having frequency p at time t+dt Probability of moving from p-ε to p

36. More proper derivation Looking backward in time, as or coalescence gives Kolmogorov backward equation Sign of directional term flips because now going backwards in times and is time reversible. Non-directional term does not flip sign because non-reversible.

37. Probability of Fixation Solve equation at and impose boundary condition for p 0 =0 and p 0 =1. Probability of Fixation of mutants

38. Investigate some limits 1.Large population, strong selection: e -2Nsp u(p 0 )~1 (guaranteed to fix) 2. Large population, really weak selection such that 2Nsp 0 <<1: e -2Nsp ~1-2Nsp 0 -> u(p 0 )~2Nsp 0 When one mutant, p 0 =1/N,and u(1/N)~2s (fixation probability increases linearly with s) 3. Under very weak selection (s->0): e -2Nsp ~1-2Nsp 0 -> u(p 0 )~ p 0, When one mutant, p 0 =1/N,and u(1/N)~1/N (same as for pure drift) All limits check out.

39. 5. Economics Black-Scholes model Price of stock is like position space (physics) or frequency space (population genetics). Directional force--general increase in worth of the market, represented by interest rate. Nondirectional force--random forces in market. Individual stocks or groups of stock will wander randomly in price. (major insight of this model! Also, because it shows value of volatility and how to make money from it.) Price of stock is like position space (physics) or frequency space (population genetics). Directional force--general increase in worth of the market, represented by interest rate. Nondirectional force--random forces in market. Individual stocks or groups of stock will wander randomly in price. (major insight of this model! Also, because it shows value of volatility and how to make money from it.)

40. Assumptions of Black-Scholes 1.Price follows Brownian motion 2.It is possible to short sell stock (options) 3.No arbitrage is possible (no asymmetry of which to take advantage) 4.Trading is continuous 5.No transactions costs or taxes 6.Stock’s price is continuous and can be arbitrarily small 7.Risk-free interest rate is constant 1.Price follows Brownian motion 2.It is possible to short sell stock (options) 3.No arbitrage is possible (no asymmetry of which to take advantage) 4.Trading is continuous 5.No transactions costs or taxes 6.Stock’s price is continuous and can be arbitrarily small 7.Risk-free interest rate is constant

41. Results from Black-Scholes Provides method for calculating fair cost of an option. Provides method for hedging “bets” and getting risk- free investment that allows one to make money according to the overall growth of the market. Provides method for calculating fair cost of an option. Provides method for hedging “bets” and getting risk- free investment that allows one to make money according to the overall growth of the market.

42. Black-Scholes PDE S-stock price V-option cost r-interest rate  -variance of random process S-stock price V-option cost r-interest rate  -variance of random process

43. Impact of this work One equation, similar concepts, applications to multiple fields with its own set of insights 1.Fourier developed the heat equation 2.Brownian motion, Einstein’s greatest achievement? 3.Applied in cosmology, particle physics, etc. 4.Huge advance in population genetics, used to study molecular motors, and lots of intracellular processes 5.1997 Nobel prize in economics for Black- Scholes One equation, similar concepts, applications to multiple fields with its own set of insights 1.Fourier developed the heat equation 2.Brownian motion, Einstein’s greatest achievement? 3.Applied in cosmology, particle physics, etc. 4.Huge advance in population genetics, used to study molecular motors, and lots of intracellular processes 5.1997 Nobel prize in economics for Black- Scholes

44. Conclusions 1.Diffusion equations describe directional and nondirectional forces. (Could also have forces on higher-order moments by extending this.) 2. Because of generality of 1, we can apply them to many different types of problems in many different fields. (Multidisciplinary) 3. Diffusion equations have already proved very useful in physics, biology, and economics. 4. Examples of two types of multidisciplinary science: a. Results from one field directly place constraints on or are utilized by agents in the other field. b. By the correct choice of analogy between fields, mathematical treatments and results can be used to draw new conclusions and insights within another field. 1.Diffusion equations describe directional and nondirectional forces. (Could also have forces on higher-order moments by extending this.) 2. Because of generality of 1, we can apply them to many different types of problems in many different fields. (Multidisciplinary) 3. Diffusion equations have already proved very useful in physics, biology, and economics. 4. Examples of two types of multidisciplinary science: a. Results from one field directly place constraints on or are utilized by agents in the other field. b. By the correct choice of analogy between fields, mathematical treatments and results can be used to draw new conclusions and insights within another field.