I wrote a small Python function using random.choice to generate sequences from the four DNA bases, G, C, A, and T:
from random import choice
symbols = ["G", "C", "A", "T"]
def strGen(x):
s = ""
for n in range(x):
s += choice(symbols)
return s
However, six consecutive calls to strGen(10) happened to produce strings beginning with A. Since each base should have a 25% chance of being selected, this seems unusually unlikely. Is something wrong with Python's random number generator, or is this just an example of randomness occasionally looking non-random?
3 Answers
For biological accuracy, this function is generating a possible sequence for one DNA strand. DNA bases pair with bases on the opposite strand—A with T and C with G—but a sequence is usually written for only one strand because the complementary strand can be inferred. RNA would use U instead of T, so including T does make this DNA rather than RNA.
The first result is also what makes the pattern seem notable. If the first sequence had started with T instead, you might have asked why all the following ones started with T. Once a particular base catches your attention, it is easy to overlook how many other runs would have seemed interesting in a different way. Python's random.choice is not being reseeded on every call here.
Nothing is wrong with the function. Each call has a 1-in-4 chance of starting with A, so six A-starting results in a row have probability (1/4)^6, or about 0.024%. That's rare, but it will happen occasionally. Generate hundreds or thousands of sequences and count the first letters to see the frequencies converge toward roughly 25% each.
Exactly. A sequence such as 5'-ATGGCTAGTAAG-3' implies the complementary strand 3'-TACCGATCATTC-5'; you do not need to include both strands when writing the sequence.