Translated with AI assistance
Chinese scientists have developed a system called DNA-SaM, which makes it possible to convert digital information into DNA sequences suitable for physical recording significantly faster. The very idea of using DNA as a data storage medium has been explored for many years; the novelty of the Chinese development lies in its encoding speed and its ability to automatically account for the biological constraints of the molecule. The study was published on October 19 in Nature Communications.
The principle is similar to conventional digital memory: only instead of zeros and ones, information is recorded using a sequence of the four DNA «letters» – A, T, G, and C. The potential density of such memory is enormous. The developers of DNA-SaM achieved a rate of 1.88 bits per nucleotide, or about 270 petabytes per gram of DNA. Forbes compares this to the storage capacity of more than 100,000 modern laptops.
Simply converting a regular file into a sequence of A, T, G, and C is not enough. Some nucleotide combinations are difficult to synthesize or read; long repetitive regions increase the likelihood of errors. If the data is stored in a living cell, the artificial sequence must also avoid accidentally activating genes or disrupting the organism’s vital functions. DNA-SaM automatically excludes such problematic combinations.
The main advantage of the new system is the speed of computational encoding. According to the authors’ estimates, traditional methods would take several years to encode 1 TB of data on a standard computer. In simulations, DNA-SaM handled this task approximately 10¹⁶ times faster than some existing algorithms.
To test the system in practice, the researchers encoded three files into 519 DNA fragments. Among them were two images of 20.2 KB and 1.06 MB, and Ernest Hemingway’s «A Very Short Story,» which was 3288 bytes in size.
The story was inserted into the DNA of living E. coli bacteria. The bacteria multiplied and, along with their own genetic information, copied the recorded file. Scientists checked the data in subsequent generations and were able to fully reconstruct the text even after 100 reproductions of the carrier cell.
DNA is unlikely to replace SSDs or smartphone memory in the foreseeable future. The main potential area of application is «cold» archives: enormous volumes of information that must be stored for decades but are rarely accessed. Current research into DNA storage specifically views long-term archiving as one of the most realistic initial applications of the technology.
This could include government and corporate archives, scientific data, genomic and medical databases, satellite observations, libraries, museum collections, and backups of large data sets. The advantages of DNA here are its high density, durability, and the fact that there is no need to continuously supply power to the storage medium.
The rise of artificial intelligence makes this challenge particularly evident: AI models, scientific computing, video, and industrial systems are generating ever more data, a significant portion of which must be retained even though constant fast access to it is not required. Therefore, over time DNA could become an additional archival layer in data center infrastructure rather than a replacement for RAM or flash memory. This scenario aligns with the current direction of research into molecular storage.
DNA-SaM solves the problem of computational encoding, but the most challenging part of the technology still lies ahead. To actually write a file, the required DNA molecule must be physically synthesized, and to read it — sequenced. Today, these processes are neither fast nor cheap. It is precisely synthesis and reading that currently represent one of the main obstacles to the commercialization of DNA storage.
Therefore, the Chinese development does not yet mean the arrival of a «DNA flash drive.» It merely removes one of the barriers on the path to scaling up the technology. If the industry manages to reduce the cost and increase the speed of DNA synthesis and sequencing just as dramatically, molecular storage devices could claim a practical role in storing massive data archives.
