DNA data storage using epigenetic modifications

DNA has tremendous potential as a data storage medium, but the process of synthesizing DNA from scratch is time-consuming. It has to be done one nucleotide at a time in a specific sequence. New research now demonstrates an alternative approach, a synthesis-free method that uses a universal DNA template to encode data. (Nature 2024, DOI:10.1038/s41586-024-08040-5)

The traditional way is to convert binary data into the DNA nucleotide bases adenine, guanine, cytosine, and thymine, and then synthesize that DNA sequence. What molecular engineer Hao Yan of Arizona State University and colleagues in China and Germany propose, instead, is to record the streams of zeros and ones directly onto the DNA itself by selective methylation. In other words, if the base is methylated, it is coded as a 1; if not, it is coded as a 0. “It’s not the [DNA] sequence itself that carries information,” says Long Qian, a researcher at Peking University who is a coauthor of the study, “rather, it’s all in the methylation layer.”

The method is based on a naturally occurring process called epigenetic modification—chemical changes to the DNA that affect its expression. One way this happens is by methylation, that is, the addition of methyl groups to the cytosine base of the DNA.

The researchers created a programmable self-assembly “typesetting” method—much like using movable type for printing on paper—to store data. First, they fabricated a universal single-strand DNA, as well as library of hundreds of short, single DNA strands called bricks. Each brick is composed of 24 nucleotides, including cytosine sites for methylation.

The DNA bricks were then typeset onto the DNA template in a fixed sequence, each brick containing a single bit of information (called an epi-bit) in the form of a 0 or 1. The DNA was then treated with the enzyme methyltransferase, which recognizes the methylated cytosine and copies the methylation information to the DNA template. This process is similar to how digital data is recorded on a hard drive by changing the direction of magnetization.

The information thus encoded was retrieved via nanopore sequencing, a technology that allows direct, real-time reading of DNA or RNA fragments and their methylation status. The researchers used the method to store and retrieve an image of an antique tiger artwork from China and a photo of a panda, about 269,337 bits of data in total.

To demonstrate the accessibility and robustness of the data-writing approach, the researchers asked 60 volunteers without biology laboratory experience to use the platform to save and retrieve their data. “DNA storage by de novo synthesis right now has to be done on a synthesis platform,” Qian says. “But our strategy is so simple, all you need to do is to add designated reagents into the same tube. Everyone can do this with their own hands.”

The biggest advantage of the method is the ability to write to parallel bits simultaneously, she says. Adding the nucleotides one by one for de novo synthesis requires between 4 and 20 min. “In our experiments, 350 bits are simultaneously copied to the template DNA strand,” Qian says.

The researchers are working on a number of improvements, including scaling up the technology, increasing data-writing speeds, and also bringing down costs. They would also like to incorporate more base modifications into the system, instead of just cytosine, to increase data density.

Dina Zielinski, a scientist at WhiteLab Genomics in Paris who wasn’t involved in the study, says: “This method has the potential to improve both speed and cost of scalable DNA storage systems.” There is room for improvement, but it’s an impressive proof of concept for overcoming the limits of de novo DNA synthesis, she adds.—.