1

I have converted 4 characters ATCG in to binary format i.e 

00 replacing A 
11 replacing T 
10 replacing C 
01 replacing G

So a character string 

AGAGAGAGTGATAGA

after conversion will look like

001000100010001011100011001000

Once I get this value I convert this binary in to its corresponding integer i.e 

143177928.

The issue is, when I want to get back to binary again, it gives me 

0b1000100010001011100011001000

which is not the correct representation of original character string because it omits all the zeros from far left after 1.

So I have written a method implementing binary conversion and I know how long the binary string should be. So in the end I just remove 0b from the returned binary and append 0s in the far left i.e 

#zeros = length of original binary - length of returned binary (0b removed)

Is there any better way of doing this conversion??

I am coding this in python.

cs95
  • 379,657
  • 97
  • 704
  • 746
JstRoRR
  • 3,693
  • 2
  • 19
  • 20
  • yes, the problem is with your encoding scheme... any number of leading zeros is the same binary number. you can only 'zero pad' the number to a fixed length in a string representation – Anentropic Jul 13 '17 at 14:03
  • https://stackoverflow.com/questions/3252528/converting-a-number-to-binary-with-a-fixed-length – yinnonsanders Jul 13 '17 at 14:31

2 Answers2

2

You can append a flag bit after the MSB to protect all the leading zeros.

Step 1: Conversion

Add a single "flag" bit at the end and convert your bit string. 

In [6]: converted_str = '001000100010001011100011001000'

In [9]: num = int('1' + converted_str, 2)

In [10]: num
Out[10]: 1216919752

Step 2: Re-conversion

Use the format method to convert your number back to a bit string, while stripping off the first "flag" bit.

In [12]: reconverted_str = format(num, 'b')[1:]

In [13]: reconverted_str
Out[13]: '001000100010001011100011001000'
cs95
  • 379,657
  • 97
  • 704
  • 746
0

Use '{0:0{1}b}'.format(num, num_digits)

This will add leading 0's until the number is num_digits. The 'b' specifies that num should be converted to binary.

yinnonsanders
  • 1,831
  • 11
  • 28
  • this was in my original method from previous stack suggestions. but I think adding a single bit initially and stripping it in the end is more efficient. thanks for your suggestion anyways. – JstRoRR Jul 13 '17 at 16:46
  • More efficient in what way? – yinnonsanders Jul 13 '17 at 16:52
  • that way I am not dependent on the length of string to figure out how many zeros I want. With above method I know I have to add and strip a single bit so I am kind of avoiding variables for length of string and leading zeros. – JstRoRR Jul 14 '17 at 10:10