ResearchMay 20, 20260 views

Zero-shot de novo peptide sequencing with open posttranslational modification discovery.

Forget waiting for a database update or sifting through endless candidate lists. The RNovA algorithm just changed the game for de novo peptide sequencing. This new transformer-based tool, built by teams from Waterloo, Tsinghua, and several biotech companies, lets researchers identify peptides and their posttranslational modifications (PTMs) in a true zero-shot setting. No retraining. No need to know what you’re looking for ahead of time.

Nat Biotechnol

by Mao Z, Peng C, Chen Y et al.

“Zero-shot de novo peptide sequencing with open posttranslational modification discovery. Mao Z(1), Peng C(#)(2)(3), Chen Y(#)(4), Wu P(#)(2), Zhang Q(#)(1), Yu Y(#)(1), Zhang R(#)(1), Xin L(5), Shan B(5), Deng H(6)(7), Li M(8)(9). Author information: (1)David R. Cheriton School of Computer Science, University of Waterloo, Waterloo, Ontario, Canada. (2)Baizhen Biotechnologies, Wuhan, China. (3)Wuhan Biobank, Wuhan, China. (4)School of Life Science, Tsinghua University, Beijing, China. (5)Bioinformatics Solutions, Waterloo, Ontario, Canada. (6)School of Life Science, Tsinghua University, Beijing, China. dht@mail.tsinghua.edu.cn. (7)Zhejiang Key Laboratory of Multiomics and Molecular Enzymology, Yangtze Delta Region Institute of Tsinghua University, Jiaxing, China. dht@mail.tsinghua.edu.cn. (8)David R. Cheriton School of Computer Science, University of Waterloo, Waterloo, Ontario, Canada. mli@uwaterloo.ca. (9)Central China Research Institute of Artificial Intelligence, Zhengzhou, China. mli@uwaterloo.ca. (#)Contributed equally De novo peptide sequencing directly infers sequences from mass spectrometry data without relying on protein databases. Although recent deep learning models can also identify posttranslational modifications (PTMs), they require labeled training data for this task. Here we introduce rotary positional embedding-enhanced de novo sequencing algorithm (RNovA), a transformer-based de novo sequencing algorithm enhanced with relative positional embeddings and a reinforcement-learning-style sequential decision framework. RNovA enables open PTM discovery in a zero-shot setting-without retraining or a predefined list of candidate residues-while maintaining state-of-the-art performance on standard benchmarks. Demonstrating this capability, we successfully identified peptides modified by kynurenine-an uncommon and biologically relevant PTM-in clinical samples from patients with RA and validated this discovery with synthetically synthesized reference peptides. Furthermore, we demonstrated open de novo PTM discovery by analyzing the bacterial strain A1232E, which lacks a reference proteome, and detected an unannotated glutamic acid modification. RNovA enables exploration of previously inaccessible regions of the proteome, including peptides with unexpected or unannotated modifications. © 2026. The Author(s), under exclusive licence to Springer Nature America, Inc. Conflict of interest statement: Competing interests: L.X. and B.S. are employees of Bioinformatics Solutions. C.P. and P.W. are employees of Baizhen Biotechnologies. The other authors declare no competing interests.”

Here’s why this matters: Traditional deep learning models for peptide sequencing need labeled data to spot rare or novel PTMs. RNovA skips that step. Its secret weapon is rotary positional embeddings and a reinforcement learning-inspired approach. Translation: the AI understands sequence patterns and modifications better, even when they’re completely new.

The team put RNovA to the test on real clinical samples. It nailed the identification of peptides modified by kynurenine—an unusual PTM relevant in autoimmune disease research. They even confirmed the results with synthetic references. Then, they analyzed a bacterial strain with no reference proteome and picked up an unannotated glutamic acid modification. This opens up unexplored corners of the proteome to researchers everywhere.

Key takeaway:

•

Open PTM discovery is now possible—zero-shot, no retraining, no hand-holding.

•

RNovA keeps pace with state-of-the-art models for standard benchmarks, so there’s no tradeoff for flexibility.

•

Uncommon and unexpected PTMs can be identified directly from mass spec data.

Curious about the broader impact? Check out the peptide research index for more on how these algorithms fuel discovery. This is a leap for anyone working with mass spectrometry, bioinformatics, or new peptide targets. Zero-shot sequencing is here, and it’s not slowing down.

Back to News Read original on Nat Biotechnol

Zero-shot de novo peptide sequencing with open posttranslational modification discovery.

Related Reading

The STRIDE Trial and Semaglutide: Implications for Clinical Vascular Practice.

Structure-Based Adaptation of a SARS-CoV-2 Neutralizing Peptide to New Virus Variants.

Combining an α(v)β(6)-Targeted (177)Lu-Based Peptide Receptor Radionuclide Therapy with Olaparib to Boost Therapeutic Efficacy in Pancreatic Cancer.