We’re familiar with the term virus referring to both biological infections and to similar infections meant to affect computers. But researchers at the University of Washington have discovered a new and surprising conflation of these two ideas: the possibility of encoding a computer virus into a strand of human DNA. We know how it feels to catch a cold, but how does it feel to catch a malware?
A conceptually similar study in the past has involved implanting a human with a microchip containing the virus, but this new research, led by computer science professor Tadayashi Kohno, instead involved encoding patterns into DNA samples. When run through a DNA sequencer, these patterns were designed to create a ‘buffer overflow’ and gain control of the computer system.
The scientists discovered, however, that the process raised unexpected complications when attempted through the medium of DNA. DNA sequencers mix different chemicals with the basic units of DNA: A, T, G, and C. The chemicals interact differently with each of these four bases, and produce different colours of light which allow the sequencer to identify the DNA’s pattern. The research was very complicated, but its goal was essentially to encode a virus into these A, T, G and C units.
The issues Kohno’s team ran into were related to this process of translation. The DNA, as well as fulfilling its function as a computer virus, also had to function as a stable DNA sequence. The researchers quickly found that the sequences required to create the buffer overflow were not necessarily conducive to the development of stable DNA. In other words, the DNA had to speak two languages at once.
The virus required the same strings of code to repeat over and over, which could cause the DNA to fold in on itself. Stable DNA also requires a certain proportion of As and Ts to Gs and Cs – again, not necessarily the same proportion required for the virus to function. Finally, while DNA sequencing can easily be done back-to-front, computer code can only execute when read in the right order. The research report suggests an amusing solution to this problem: writing the DNA sequence as a palindrome, so that it would read the same forwards or backwards.
When digital and biological collide
The team was eventually able to achieve some results – at least in attacking the modified version of the DNA sequencer program that was used in the tests. The research suggests that such biological hacking techniques are highly impractical and unthreatening for now. But Seth Shipman, Harvard researcher whose own work involved storing a GIF inside bacteria, commented that “down the line, when more information is stored in DNA and it’s being input and sequenced constantly, we’ll be glad we started thinking about these things”.
What will a future in which biomedicine and computer science collide look like? And will we infect our computers by coughing at them? We are used to protect ourselves against infections from tainted foods, animals and other people. And in the digital realm, we are used to downloading antivirus software to protect our computers from similar infections. But in our next nature, which is wild and unpredictable as ever, we might have to get used to protecting ourselves against computer viruses, and protect our computers against influenza; and no handkerchief can top that!