Proteins, made of amino acids strung together in peptide chains, carry out many essential tasks in living cells, yet determining their exact sequences has remained a persistent challenge. While DNA sequencing has become fast and inexpensive, comparable tools for protein sequencing have lagged. A team led by bioengineers at Stanford University has now described an approach that effectively “runs the natural process—in reverse” by converting protein sequence information into DNA, allowing scientists to take advantage of existing high-throughput DNA sequencing platforms. Their work, published in Nature Biotechnology, outlines a chemical process that increases the scale and sensitivity of protein analysis.
DNA provides the instructions for building cellular machinery, but proteins are the machinery itself, folding into complex shapes that control growth, communication, immune responses, and disease processes. Sequencing proteins is more difficult than sequencing DNA because proteins are constructed from 20 amino acids instead of four bases and the amino acids are smaller than nucleic acids, making them harder to distinguish reliably. According to first author Liwei Zheng, current mass spectrometry methods typically detect only about a million molecules out of billions analyzed. By contrast, the new method may allow detection of up to a thousand times more molecules from the same sample, potentially revealing rare protein types that are hidden in bulk analyses.
The technique works by chemically tagging individual amino acids within a peptide using DNA barcodes that are unique to each peptide molecule. Antibodies and cyclic barcodes made of synthetic DNA help both encode each peptide’s identity and mark its position. Once these tags are in place, conventional DNA sequencing can read out the amino acid composition and its location in the peptide chain, effectively turning protein information into DNA sequence data. Senior author H. Tom Soh highlighted that this strategy enables sequencing of individual proteins at the single-molecule level, using very small amounts of material, which could extend down to individual cells.
This level of sensitivity may support detailed studies of cellular diversity, including why seemingly similar cells respond differently to diseases such as cancer. For example, in immunotherapy settings like CAR-T cell treatment, comparing immune cells that respond well to therapy with those that do not at the single-protein level could clarify why certain treatments succeed only in specific cancers or cell types. The approach has already drawn commercial interest, with plans to develop a user-friendly instrument that allows researchers to load a sample, start the process, and obtain results in a manner similar to running a DNA sequencer today.
Although still in early development and requiring further optimization, the method is designed to scale to billions of protein molecules from thousands of cells. Zheng noted that once protein information is converted into DNA, it becomes possible to use established DNA-handling tools—such as copying and lengthening DNA—to process protein sequences.
Publication details: Zheng, L., Sun, Y., Hein, L.A. et al. Single-molecule peptide sequencing through reverse translation of peptides into DNA. Nat Biotechnol (2026). https://doi.org/10.1038/s41587-026-03061-z
DNA provides the instructions for building cellular machinery, but proteins are the machinery itself, folding into complex shapes that control growth, communication, immune responses, and disease processes. Sequencing proteins is more difficult than sequencing DNA because proteins are constructed from 20 amino acids instead of four bases and the amino acids are smaller than nucleic acids, making them harder to distinguish reliably. According to first author Liwei Zheng, current mass spectrometry methods typically detect only about a million molecules out of billions analyzed. By contrast, the new method may allow detection of up to a thousand times more molecules from the same sample, potentially revealing rare protein types that are hidden in bulk analyses.
The technique works by chemically tagging individual amino acids within a peptide using DNA barcodes that are unique to each peptide molecule. Antibodies and cyclic barcodes made of synthetic DNA help both encode each peptide’s identity and mark its position. Once these tags are in place, conventional DNA sequencing can read out the amino acid composition and its location in the peptide chain, effectively turning protein information into DNA sequence data. Senior author H. Tom Soh highlighted that this strategy enables sequencing of individual proteins at the single-molecule level, using very small amounts of material, which could extend down to individual cells.
This level of sensitivity may support detailed studies of cellular diversity, including why seemingly similar cells respond differently to diseases such as cancer. For example, in immunotherapy settings like CAR-T cell treatment, comparing immune cells that respond well to therapy with those that do not at the single-protein level could clarify why certain treatments succeed only in specific cancers or cell types. The approach has already drawn commercial interest, with plans to develop a user-friendly instrument that allows researchers to load a sample, start the process, and obtain results in a manner similar to running a DNA sequencer today.
Although still in early development and requiring further optimization, the method is designed to scale to billions of protein molecules from thousands of cells. Zheng noted that once protein information is converted into DNA, it becomes possible to use established DNA-handling tools—such as copying and lengthening DNA—to process protein sequences.
Publication details: Zheng, L., Sun, Y., Hein, L.A. et al. Single-molecule peptide sequencing through reverse translation of peptides into DNA. Nat Biotechnol (2026). https://doi.org/10.1038/s41587-026-03061-z
