The invention of the printing press and movable type—metal letters that can be arranged and inked—led to the Renaissance and an explosion of information that continues to this day. Now, researchers report applying the concept of movable type at the molecular level to dramatically speed up the ability to encode data in strands of DNA, an incredibly high-density medium for storing information. Although only demonstrated in the lab so far, the new approach, reported today in Nature, could energize the emerging DNA data storage industry by making it cost effective to archive vital information for decades and beyond, independent researchers say.
“It’s a really nice proof of concept and a significant improvement over previous DNA data storage approaches,” says Kun Zhang, a genomics expert at Altos Labs. “It gets around a barrier of DNA data storage that requires synthesizing DNA from scratch,” adds Jeff Nivala, a biophysicist at the University of Washington.
The allure of DNA data storage is immense: A single gram of DNA can store up to 215 petabytes of data, enough to store 10 million hours of high-definition video. At that rate, a few pickup trucks worth of DNA could store all the data humanity has ever recorded. And unlike conventional electronic hard drives, which degrade in years or decades, DNA can last for millennia.
Moreover, reading out data encoded in DNA’s four-letter alphabet is straightforward and relatively fast these days with DNA sequencing machines. The problem is writing the data, which typically requires synthesizing custom strands of DNA one letter at a time. Today’s fastest DNA writers can synthesize about 320 million bytes of DNA data per day. At this speed, writing a single gram’s worth of DNA would take nearly 2 million years. “It’s unaffordable compared to hard drives because the writing speed is quite slow,” says Long Qian, a computational biologist at Peking University.
To speed things up, Qian and her colleagues sought inspiration from movable type, originally invented in China around 1040 C.E., some 400 years before Johannes Gutenberg’s printing press, using porcelain instead of metal. To serve as the “paper,” they synthesized long standardized, template pieces of single-stranded DNA. For the “type,” they synthesized hundreds of short single-stranded DNA “bricks,” each 24 bases long, with a sequence designed to bind to a specific region along the DNA template.
The researchers then turned to a natural process in cells called methylation to encode the bricks with either a digital 0 or 1. In the body, cells attach methyl groups—one carbon and three hydrogen atoms—to specific DNA sequences to signal which genes should be expressed and silenced in different tissues.
For DNA storage, Qian and her colleagues added an enzyme that attached methyl groups to some of the DNA bricks (the 1s) and left others alone (the 0s). Like Renaissance typesetters, they then selected the bricks that would align to the template with the proper 1s and 0s to encode whatever digital file they wanted. Placed in solution, the bricks quickly found and connected to their corresponding sequences on the template.
Finally, the researchers had to “ink” the type and print to the paper. They added an enzyme known as a methyltransferase, which copied all the methyl groups on the bricks to the adjacent location on the template strand, which could then be read out by a commercial DNA sequencing machine. They demonstrated writing and reading files containing nearly 270,000 bits, enough to code high-resolution images such as a tiger and giant panda.
Qian says the cost of writing data with the approach is, at the moment, about $0.003 per bit. Although that’s higher than what DNA synthesis companies charge for writing each new letter of DNA, Qian believes a commercialized operation would reduce costs by using fewer reagents than her lab.
But Qian says the study already demonstrates a speed-up in writing data into DNA. Qian estimates a commercial version of the new approach could reach speeds of up to 2 terabytes per day, a 6000-fold increase over today’s best commercial DNA synthesizers. Qian and her colleagues are now looking to decorate the DNA templates with other chemical markers besides methyl groups to encode even more data per strand and speed things up even more. If successful, movable type may not just be one of antiquity’s breakthroughs, but one for the future, too.
More: https://www.science.org/content/article/dna-printing-press-could-quickly-store-mountains-data
