This post is the first in a series exploring whether it is possible to replace our current computers with biological computers. Of course, you could argue that every living creature is a biological computer. My question is whether we can take the tools our bodies use to store, transmit and decode information and produce more efficient, smaller computers.
First I want to look at information storage. DNA has a proven track record at efficient information storage. We have a full copy of all 3 billion base pairs that make up our DNA packed into the tiny nucleus in every cell in our bodies. In 2012, scientists used DNA to store over 750 kilobytes of data. They turned five different computer files–a text file containing all of Shakespeare’s sonnets, an MP3 of Martin Luther King’s “I have a dream” speech, a pdf copy of the famous Watson and Crick paper on the structure of DNA, and a JPEG photo–into tiny white specks of DNA that could barely be seen with the naked eye.
To do this, they first recoded the information from bytes (a series of eight bits) into five DNA bases. Computer information is currently stored in bits. A bit is a binary digit with only two possible values (commonly represented as 0 and 1 or hi and lo). DNA is encoded by four different bases: A, T, C and G. Therefore DNA has the potential to store information more efficiently than bits. The recent invention of synthetic DNA will increase this ability even further.
When making the DNA code, the scientists made sure it did not include series of the same base in a row eg AAAA or TTT because this can lead to errors when the DNA is read. The theoretical code was then synthesised into real DNA in a lab. It is expensive and difficult to make long sequences of DNA, so instead the scientists broke up the encoded information into pieces 117 bases long. The small pieces were designed to overlap. This redundancy also reduced the risk of reading error as a mistake in the code would have to be in 4 different DNA pieces to be decoded at the other end. Indexing information was included in each 117 base segment so it would be clear where that segment fit in relation to the rest of the code. Since the DNA segments in this experiment were all the same length and had no runs of the same base, they were obviously different from DNA found in natural organisms.
The DNA was then sent from the US to a lab in Germany where it was sequenced. This sequence was used to reconstruct the original computer files with 100% accuracy. This was despite a sequencing error rate of 1:500 bases. The overlapping sequences allowed the errors to be picked up and those DNA strings were omitted from further analysis.
While this sounds like a cool experiment, what is the point? Well, one of the major benefits that DNA has over standard computer information storage is that DNA can be stable for tens of thousands of years, if not longer. DNA has been successfully sequenced from a long dead woolly mammoth. Current computer technology can only reliably store information for 10 years.
For now, it is still expensive and slow to write DNA. However, the costs of reading DNA have dropped dramatically in the last few years. It now costs only $1000 to sequence the entire human genome. Until it also becomes cheaper and easier to write DNA, it is only feasible to use DNA for archiving rather than storage of frequently accessed information.
Read the other posts in the bio-computing series here:
The Scientific Papers: