As if you aren't seeing GIFs everywhere already, now scientists are adding them to the DNA of bacteria, too.

Adding data to DNA isn't anything new. Last year, Microsoft embedded an OK Go music video, Leo Tolstoy's "War and Peace" and the Universal Declaration of Human Rights to a bit of synthetic DNA that was "much smaller than the tip of a pencil." It was an impressive way to store and retrieve 200 megabytes of data.

Now, a team of researchers out of Harvard have used the gene editing tool CRISPR/Cas to insert five frames of a horse galloping and a picture of a human hand into the DNA of living E. coli. The study, published in Nature, further pushes the boundaries of using DNA to store information.

This wasn't 200 megabytes worth of data, of course. The GIF, for instance, was only 36 by 26 pixels large and had a file size of 2.6 kilobytes. The picture of the hand was a 4-color 56-by-56-pixel image with a file size of 706 bytes. For reference, the picture at the top of this file is 653 by 415 pixels and is almost 109 kilobytes in file size. The size was not the thing. The boundary-pushing comes in the fact that it was added to the DNA of living bacteria, not synthetic DNA.

Data retrieval

In order to encode the data onto the genomes of the bacteria, researchers first had to convert the individual pixels of the GIF and the image into nucleotides, the basic foundation for DNA. Then the researchers used the CRISPR/Cas tool to insert bits of the pixels into multiple E. coli cells.

It was not a quick process. The GIF of an Eadweard Muybridge-photographed galloping horse — Muybridge's photo experiments with horses in the late 1880s are considered a forerunner to movies, making this a rather appropriate GIF to use — took five days to transfer its five frames to the bacterial cells.

"The information is not contained in a single cell, so each individual cell may only see certain bits or pieces of the movie. So what we had to do was reconstruct the whole movie from the different pieces," co-author of the study, Seth Shipman, explained to the BBC.

"Maybe a single cell saw a few pixels from frame one and a few pixels from frame four ... so we had to look at the relation of all those pieces of information in the genomes of these living cells and say: Can we reconstruct the entire movie over time?"

The answer to Shipman's question was yes. The team sequenced the DNA of the bacteria and managed to reconstruct both the image and the GIF with 90 percent accuracy. The video above shows the GIF as it was encoded into the bacteria on the left, and on the right is the GIF that the researchers got back. It's a little spottier than the original, but it's clearly the same image.

Memories living on

OK, so great, scientists managed to splice a GIF of a 19th-century horse galloping on a small picture of a human into E. coli DNA. So what?

Well, there's the obvious data storage benefits of this, which is why researchers — and tech companies, like Microsoft — have turned to using DNA as replacements for more traditional forms of data storage. Living DNA is great for this because it can last for thousands of years and, as Yaniv Erlich, a computer scientist and biologist at Columbia, pointed out to MIT's Technology Review, some bacteria can survive in extreme conditions, including nuclear explosions and very high temperatures. So these bacteria function as natural shields to keep the data safe.

However, living DNA is harder to encode information on since cells are always moving, dividing and sometimes dying off as well. You could potentially lose a fair amount of data just due to that. And, as has been pointed out, the amount of information encoded onto the living DNA pales in comparison so far to that of the data edited into synthetic DNA. For now, synthetic DNA is probably the way of the future for storing information.

The researchers weren't primarily interested in that kind of data storage, however. They're more interested in turning the living cells into sensors that track and record information about what's happening in the cells themselves. This is why the researchers used a GIF of a horse galloping. There's a timing element involved in the retrieving and sequencing of that GIF, the same as there would be when tracking the changes in a cell. This kind of self-gathered information could be related to tracking the catalysts for cellular evolution, like the formation of neurons in the brain.