# Is there a way to measure the amount of bytes that are possible to encode in a DNA molecule?

When I saw a DNA molecule for the first time, it kinda reminded me of a hard drive. It consists of slots and there are some possible combinations for each slot; in the hard drive these possible combinations would be 0's and 1's. In DNA, these slots would be G's, A's, T's, C's.

So, is there a way to measure the amount of bytes that are encoded in a DNA molecule?

I've made this question before in another forum, but the answerer provided me only with Shannon's theorem, which is $K=L-\frac{(1-q^L)^n}{q^L}$ and told me a little about genetic redundancy. I could only search for the ammount of slots which are present in the DNA, but this genetic redundancy thing got me stuck.

• I am not sure what your question is - there are indeed four nucleotide bases in DNA, and these correspond to the "slots" you mention. You may want to have a look here: bitesizebio.com/articles/…
– blep
Mar 22, 2013 at 7:27
• This is the answer. I was just confused to the real ammount of data - because of the genetic redundancy. Mar 22, 2013 at 7:46
• You might find this thread over at skeptics interesting. Mar 22, 2013 at 16:46

## 2 Answers

Unfortunately the answer is highly dependent on what you mean. In the simplest terms, comparing it directly to how we measure data storage in digital media, the number of different states of a DNA string of length $n$ can have is simply $4^n$. A byte holds $2^8$ different states so the number of bytes in a DNA string of length $n$ is $\frac{n}{4}$. Of course, actually accessing this information would be more difficult than simply having it in a strand of DNA.

DNA in real organisms is not random, however, so sequences are not randomly distributed meaning that you could compress that information down into less bits than that, Shannon information style.

However, you can decently argue that this doesn't really tell us how much information is in real DNA because real DNA has structure that matters. It has exons and introns and promoter regions and so on. Meanwhile, the sequences in protein coding regions are a lot more important than sequences in non-coding regions for the most part. Large parts of the genome are functionally irrelevant but because of this they tend to be more random and thus have higher Shannon information.

• And then there was ENCODE's definition of function...
– blep
Mar 22, 2013 at 20:35

You might also be interested in this paper from EMBL-EBI about storing data on DNA.

Towards practical, high-capacity, low-maintenance information storage in synthesized DNA

They show they can get 757,051 bytes or a Shannon information 10 of 5.2 × 106 bits onto 153,335 strings of DNA, each comprising 117 nucleotides (nt).

George Church had a similar paper recently as well - Science