In python 2, division of integers is not true division. This will be changed in python 3. For example:

In [ ]:

1/4

Note that this isn't rounding, but truncation of the remainder.

In [ ]:

3/4

There are a few ways around this, but most convenient is to add the the following import statement. It's always a good idea to add this to your python code to avoid any surprises.

In [ ]:

from __future__ import division

In [ ]:

3/4

Genetic bottleneck programming example

This example illustrates the concept of genetic drift in a population of genotypes. Imagine we start with four genotypes, A, B, C, and D and 10,000 individuals. First, let’s define a starting population with the following genotype frequencies:

In [ ]:

genotype_frequencies = ['A'] * 5000 + ['B'] * 2500 + ['C'] * 1250 + ['D'] * 1250

This syntax may be new to you. Break this apart to individual addends to figure out what the full statement is doing.

Next, let’s define a function that will conveniently allow us to summarize this population:

In [ ]:

def summarize_composition(population):
    population_size = len(population)
    print 'A: %0.4f' % (population.count('A') / population_size)
    print 'B: %0.4f' % (population.count('B') / population_size)
    print 'C: %0.4f' % (population.count('C') / population_size)
    print 'D: %0.4f' % (population.count('D') / population_size)

You can then call this as follows:

In [ ]:

summarize_composition(genotype_frequencies)

Next we’re going to import the sample function from the random module. Given a list (population) and a number of elements (k) to select, sample randomly samples (without replacement) k elements from the list and returns those as a new list. So, if we sample the full population and summarize the genotype composition, we should get the same result – let’s test it out:

In [ ]:

from random import sample
new_genotype_frequencies = sample(genotype_frequencies,10000)
summarize_composition(new_genotype_frequencies)

Now, let’s simulate genetic drift. Imagine we have a population of organisms with the genotype frequencies represented in our genotype_frequencies list. Regardless of which of these genotypes confers the most selective advantage a random removal of a large component from the population has the ability to affect the resulting genotypic composition.

To simulate an event that randomly kills off 10% of the total population, and look at the resulting genotype composition:

In [ ]:

new_genotype_frequencies = sample(genotype_frequencies,9000)
summarize_composition(new_genotype_frequencies)

Do this a few times. You should notice that the frequencies don’t change a lot. What happens if instead of this relatively small dying off, there is a near-extinction event. Simulate an event that randomly kills off 99.9% of the population. What happens now? Run this simulation several times and explain the results of this experiment.