What's a strand of DNA but data? We often think of its units, the As, Cs, Ts, and Gs, as letters of the words in an instruction manual. But what if, instead, we think of them as biological computer bits, storing the smallest unit of information? What stops scientists from harnessing the power of those units, using the latest biological technology to treat DNA like a writable disk?
A team of biologists took advantage of the bacterial defence mechanism on which the gene editing system CRISPR is based to write an animated GIF, pixel by pixel, into a population of bacteria. But their end goal isn't to create biological DVDs. Instead, they hope to one day use this technology as a sort of biological recorder to keep track of cells or changes in the environment.
"We used images in movies to show the power of the system that we want to use to capture all sorts of biological activity into DNA," the study's first author, Seth Shipman from Harvard Medical School told Gizmodo. "The idea is to create molecular recorders that function in cells to capture things in the cell or cell environment .... so that an experimenter doesn't have to go in and disrupt the system to collect data."
Bacteria already use the "Cas" proteins associated with CRISPR to grab pieces of DNA from viruses and insert it into their genomes as a defence mechanism against future attacks. The researchers took advantage of the fact that the proteins always insert new genetic material upstream from the old genetic material. "You end up with a literal, physical record of which sequences have been put in there in what order," said Shipman. "The actual timing information is carried in the order of the sequences of the genome as you read it." So, they made movies to show off DNA as an "excellent medium for archiving data" according to the paper published today in the journal Nature.
The researchers first tested storing an image of a hand using DNA by converting colour and pixel information into sets of base pairs, those As, Ts, Cs, and Gs. This is kind of like the way that genes code for proteins in DNA, but in this custom DNA they built, the A, C, T, G sequences correspond to barcodes that determine which pixels should take on which colour. Then, they introduced their custom sequences into a population of E. coli bacteria, creating temporary pores in the cells' membranes with electric pulses. The translated data enters into the bacteria, which integrate it into their genome using their Cas proteins. The team then grows the bacterial population, which they then sequence, reading the DNA, and translate, recreating the image.
Encoding a GIF into the bacteria was a more complicated process — even though the bacteria store the data in chronological order, the GIF information gets scattered among lots of cells. The researchers can only recover the information on the ordering of frames from single cells, according to the paper. That means recreating the movie required comparing the DNA from lots of different cells. Here's the GIF before the insertion into the bacteria and the reconstruction.
There are limitations that come alongside working with a living data storage system, said Shipman. The scientists are trying to pull a lot of information out of these cells, which can lead to dead pixels or errors — they could reproduce the GIF with 90 per cent accuracy, not perfectly. Additionally, given their methods, the more cells they read back, the more accurate their GIF was; they needed to come up with a reasonable place to stop. Here's how the hand reconstruction got better over time:
Other researchers I spoke with found the research interesting. "The methods described are not that new, but it is a super-cool application," researcher Christopher Voigt from MIT told Gizmodo in an email. He recently engineered bacteria to make coloured pictures. "Here, the authors use low-cost DNA synthesis and genome editing tools to put the information from pictures and movies into the genome of a living cell. Every time the cell replicates, a new copy is made. Puts a new spin on pirated material!"
Shipman hopes to use this method in biological applications — for example, using bacteria to sense heavy metals in the environment, and keep a DNA record of those contaminants over time.
But just imagine how many bacteria it would take to recreate, say, Contagion.