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

2026-05-19 · Source: Machine learning : nature.com subject feeds · Field: Science & Research — Life Sciences & Biology, Health & Medical Research, Artificial Intelligence & Machine Learning · Depth: Expert, long

Summary

RNovA, a novel transformer-based de novo peptide sequencing algorithm, introduces rotary positional embeddings and a reinforcement-learning-style sequential decision framework to enable open posttranslational modification (PTM) discovery. This algorithm operates in a zero-shot setting, meaning it identifies PTMs without requiring labeled training data or a predefined list of candidate residues, while maintaining high performance on standard benchmarks. Demonstrating its capabilities, RNovA successfully identified kynurenine-modified peptides in clinical samples from rheumatoid arthritis (RA) patients, validating these findings with synthetically synthesized reference peptides. Furthermore, it detected an unannotated glutamic acid modification in the bacterial strain A1232E, which lacks a reference proteome. The RNovA framework, with its source code available on GitHub and datasets on Zenodo (doi.org/10.5281/zenodo.15715597) and PRIDE (PXD076296), expands access to previously unexplored regions of the proteome, including peptides with unknown modifications.

Key takeaway

For proteomics researchers investigating novel posttranslational modifications or analyzing proteomes without existing annotations, RNovA provides a significant advancement. You can now perform zero-shot PTM discovery, eliminating the need for extensive labeled training data and predefined modification lists. This capability allows you to explore previously inaccessible regions of the proteome, such as identifying kynurenine in RA patients or unannotated glutamic acid modifications in bacterial strains. Consider integrating RNovA into your workflow to broaden your discovery scope and accelerate research into unexpected protein modifications.

Key insights

RNovA enables zero-shot open PTM discovery in de novo peptide sequencing with high accuracy.

Principles

• Zero-shot PTM discovery removes labeled data dependency.
• Relative positional embeddings improve transformer sequencing.
• Reinforcement learning guides sequential peptide inference.

Method

RNovA employs a transformer architecture with rotary positional embeddings and a reinforcement-learning-style sequential decision framework to infer peptide sequences and identify PTMs from mass spectrometry data.

In practice

• Discover uncommon PTMs in clinical patient samples.
• Sequence peptides from organisms lacking reference proteomes.
• Uncover unannotated modifications in proteomic analyses.

Topics

• De Novo Peptide Sequencing
• Posttranslational Modifications
• Zero-shot Learning
• Mass Spectrometry
• Transformer Models
• Proteomics

Code references

• zqq66/RNovA

Best for: AI Scientist, Research Scientist, Domain Expert

Related on AIssential

• A large-scale unified deep learning model for peptide mass spectrum interpretation trained on multimodal data — pUniFind unifies peptide-spectrum scoring and de novo sequencing, significantly improving proteomic analysis sensitivity and modification coverage.
• Single-round evolution of RNA aptamers with GRAPE-LM — GRAPE-LM enables single-round RNA aptamer evolution using AI, outperforming multi-round human methods.
• DDA-BERT: end-to-end training for data-dependent acquisition mass spectrometry-based proteomics — DDA-BERT is a transformer-based model that significantly improves peptide identification in DDA proteomics through end-to-end deep learning.
• Probabilistic modelling of single-cell bisulfite sequencing data with MethylVI — MethylVI disentangles biological signals from technical noise in single-cell bisulfite sequencing data.
• 🔬ESM: The Bitter Lesson is Coming for Proteins - Alex Rives, BioHub — Scaling protein language models with vast, diverse data reveals emergent biological capabilities and enables programmable design.
• Rznomics Reports Interim Clinical Data for RZ-001 in Hepatocellular Carcinoma at AACR 2026 Demonstrating Encouraging Efficacy and Favorable Safety Profile — RZ-001, an RNA editing therapy, shows promising efficacy and safety for hepatocellular carcinoma.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine learning : nature.com subject feeds.