Generating random sequences of DNA

Question

I am trying to generate random sequences of DNA in python using random numbers and random strings. But I am getting only one string as my output. For example: If I give DNA of length 5 (String(5)), I should get an output "CTGAT". Similarly if I give String(4) it should give me "CTGT". But I am getting "G" or "C" or "T" or "A"; i.e. only a single string each time. Could anyone please help me with this??

I tried the following code:

from random import choice
def String(length):

   DNA=""
   for count in range(length):
      DNA+=choice("CGTA")
      return DNA

score 10 · Answer 1 · answered Jan 18 '14 at 15:14

10

I'd generate the string all in one go, rather than build it up. Unless Python's being clever and optimising the string additions, it'll reduce the runtime complexity from quadratic to linear.

import random

def DNA(length):
    return ''.join(random.choice('CGTA') for _ in xrange(length))

print DNA(5)

answered Jan 18 '14 at 15:14

Paul Hankin

54,811
11
92
118

Whoever voted down should have left a note (such as "the question is about understanding code blocks and indentation"). Using `str.join()` is _much_ faster than adding letters (DNA bases) one by one. – lgautier Jan 18 '14 at 15:22
instead of random.choice('CGTA') you can put the frequency for each base down, like this: random.choice("A"*5+"C"*4+"G"*3+"T"*10). Replace the numbers I gave by the real relative numbers you need to use. – Stefan Gruenwald Aug 03 '14 at 16:07

score 7 · Accepted Answer · edited Jan 18 '14 at 17:00

You return too quickly:

from random import choice
def String(length):

   DNA=""
   for count in range(length):
      DNA+=choice("CGTA")
      return DNA

If your return statement is inside the for loop, you will only iterate once --- you will exit out of the function with the return.

From the Python Documentation on return statements: "return leaves the current function call with the expression list (or None) as return value."

So, put the return at the end of your function:

def String(length):

       DNA=""
       for count in range(length):
          DNA+=choice("CGTA")
       return DNA

EDIT: Here's a weighted choice method (it will only work for strings currently, since it uses string repetition).

def weightedchoice(items): # this doesn't require the numbers to add up to 100
    return choice("".join(x * y for x, y in items))

Then, you want to call weightedchoice instead of choice in your loop:

DNA+=weightedchoice([("C", 10], ("G", 20), ("A", 40"), ("T", 30)])

Thanks a lot. It is working now. Can i add a distribution which is a list of tuples for example String(10,[("a",20),("b",60),("c",20)]) where the probabilities sum to 100(20+60+20) — Rachel, Jan 18 '14 at 15:22

score 1 · Answer 3 · answered Jun 28 '21 at 15:57

1

Perhaps numpy works faster thanks to vectorization?:

import numpy as np
seq_length = 100
my_seq = ''.join(np.random.choice(('C','G','T','A'), seq_length ))

answered Jun 28 '21 at 15:57

porygon-tech

51
2

score 0 · Answer 4 · answered Aug 29 '14 at 05:37

I've upgraded the code to provide GC percent distribution from 0 to 100%. The code above always produces a 50% distribution.

The actg_distribution string can be any length of an existing DNA sequence of known GC percent. GC percent of a certain range is a common use case.


import random

# Return random CGTA sequences, set minimum = maximum to get a specified length.
def random_length_dnasequence(minimum=25, maximum=10000, actg_distribution=None):
    if (minimum == maximum):
        length = minimum
    else:
        length = random.randint(minimum, maximum)
    if (actg_distribution == None):
        actg_distribution = ''.join(random.choice('cgta') for _x in xrange(7))

    return ''.join(random.choice(actg_distribution) for _x in xrange(length))


def random_dnasequence(length, actg_distribution=None):
    return random_length_dnasequence(length, length, actg_distribution)

score 0 · Answer 5 · answered Oct 09 '19 at 08:47

Fast function for python 3.6 using random.choices

import random

def string(length=int(), letters="CGTA"):
        #slower 0.05s for 20000 nt
#     dna =""
#     for count in range(length):
#         dna+=choice("CGTA")
#     return dna

    #0.013s for 20000 nt
    return''.join(random.choices(letters, k=length)

Generating random sequences of DNA

5 Answers5