Harvard researchers have built a silicon chip that can write DNA, synthesizing 64 different sequences simultaneously using electricity instead of the hazardous solvents typically required for the process.
The work, published in Nature Electronics, was led by Donhee Ham, the John A. and Elizabeth S. Armstrong Professor of Engineering and Applied Sciences at Harvard's John A. Paulson School of Engineering and Applied Sciences. It marks the latest expansion of silicon chips beyond computing into biology, following earlier uses in recording neural activity and reading DNA.
Most synthetic DNA today is manufactured through phosphoramidite chemistry, a method capable of producing millions of sequences in parallel but dependent on organic solvents and specialized facilities. Researchers have long looked to enzymatic synthesis, which uses water and mimics how living cells build DNA, as a gentler alternative. Until now, enzymatic techniques had been limited to producing only about a dozen sequences at once.
The Harvard chip pushed that number to 64 simultaneous sequences, each up to 39 nucleotides long — a new benchmark for the field.
How the chip controls chemistry with current
DNA synthesis proceeds one nucleotide at a time, with a blocking group added after each step to halt growth until it is chemically removed in a process called deprotection, which requires acidic, low-pH conditions. Building many different sequences at once means lowering the pH only at chosen spots during each cycle.
The Harvard chip achieves this with 64 synthesis sites, each surrounded by two ring-shaped electrodes. An inner electrode generates protons to lower pH and trigger DNA growth at that site, while an outer electrode removes protons spreading outward, keeping the reaction confined. Repeating this cycle allows the chip to build 64 distinct DNA sequences independently.
Notably, the chip wasn't originally designed for this purpose. Jeffrey Abbott, a former PhD student in Ham's lab, had developed the underlying silicon electronics to record electrical activity in neurons. After redesigning the surface electrodes, the team realized the same precision current control could regulate the chemistry needed for DNA synthesis instead. "It worked," Ham said.
The researchers also demonstrated a potential future use by encoding a 169-byte text using the 64 synthesized sequences, pointing toward DNA-based data storage — though that application would require synthesis at a far larger scale than currently achievable.
When the team tried packing synthesis sites closer together to boost output, the attempt failed, but revealed that the chip itself was not the bottleneck. Han Sae Jung, a co-first author, said the chip successfully localized low pH as intended, but intermediate molecules generated during deprotection drifted into neighboring sites. The team says developing a more direct acid-driven deprotection chemistry is the next step needed to keep pace with the chip's capabilities.
The project involved collaborators from Harvard, the Broad Institute, DNA Script, and Pohang University of Science and Technology (POSTECH), where co-first author Woo-Bin Jung is now an assistant professor. Harvard's Office of Technology Development has filed intellectual property related to the platform.