messing around with generating strings given frequencies of substrings in python

dna_strands.py

import random

# define letters
letters = ['A', 'C', 'T', 'G']

# get a sorted list of your triplets and frequencies
# (assuming you have a way of getting these already)
# (these are for illustration)
trips = ['AAG', 'CCC', 'TTT']
freqs = [0.25, 0.70, 0.05]

def pick():

    # generate a random number between 0 and 1
    x = random.uniform(0,1)

    # create a running total as you add all frequencies
    # if the random number falls into a given frequency bucket,
    # return the triplet you want
    running_sum = 0
    for trip, freq in zip(trips, freqs):
        running_sum += freq
        if x <= running_sum:
            return trip


def assemble_string():

    # assemble a string by picking
    # 100 triplets given their frequencies
    my_string = ''
    for i in range(100):
        my_string += pick()

    return my_string


def frequency_breakdown(input_string):

    # get frequencies of all letters in a string
    frequencies = {letter: 0 for letter in letters}

    for letter in input_string:
        frequencies[letter] += 1

    return frequencies


def iterate(num_trials):

    total_frequencies = {letter: 0 for letter in letters}

    for i in range(num_trials):

        # generate string and get frequency breakdown
        s = assemble_string()
        f = frequency_breakdown(s)

        # add trials results to overall results
        for k in total_frequencies.keys():
            total_frequencies[k] += f[k]

    return total_frequencies


def convert_to_percentages(frequencies):

    total = sum(frequencies.values())

    pct_freqs = {k: float(v)/total for k,v in frequencies.iteritems()}
    return pct_freqs


# actually do the experiment
if __name__ == '__main__':

    results = iterate(10000)
    print convert_to_percentages(results)