A combined experimental and computational strategy for the analysis of ultra-short peptides (USPs)

To ensure an adequate protein supply in the context of global population growth, global agricultural production has significantly intensified. This increase in both raw and processed material production has also led to a rise in agro-industrial by-products, which some of them are particularly rich in proteins: seeds (e.g., soybean, rapeseed, peanut) contain 40–50% protein; meals from cereal and starch industries contain 25–80% protein; animal processing by-products contain 55–90% protein; and plant-derived by-products contain 25–35% protein.

These materials are generated in large volumes and are valorised for animal feed, bio-based products, biotechnology, and human food ingredients. Concomitantly, the current market size of protein hydrolysates (from dairy, plant, animal, and marine sources) is estimated at USD 3– 4 billion and expected to potentially double within the next 5–10 years according to several forecasts. In parallel, the fermented protein market is estimated at USD 1–1.3 billion.

Proteolysis, fermentation, and gastrointestinal digestion of proteins are closely interconnected processes which result in complex mixtures of free amino acids and peptides—ranging from di-, tri-, and tetrapeptides to longer chains—with substantial structural diversity and bioactivity. BioPep-UWM, one of the most widely used bioactive peptide databases, lists 5,482 bioactive peptides (as of October 2025), representing 90 distinct bioactivities. Among them, nearly 34% are ultra-short peptides (USPs define as linear di-, tri-, tetra- and pentapeptides). However, the characterization of USPs remains analytically challenging due to their physicochemical properties and several technological hurdles. We recently published a review highlighting the analytical challenges associated with the characterization of USPs (Weber et al, 2025).

This study aims to improve the characterization of proteinogenic dipeptides generated during proteolysis. A representative set of 41 peptides was:

i) Selected using 211 molecular descriptors and multivariate analysis (PCA) to ensure coverage of the physicochemical space relevant to the 400 theoretical food-derived dipeptides, and

ii) Analyzed combining orthogonal chromatographic separations (reversed-phase C18 and HILIC) coupled to tandem mass spectrometry detection to enhance selectivity. In parallel, a Python-based computational workflow was developed for high-throughput data processing.

The results highlight:

1. The strong complementarity of optimized orthogonal chromatographic modes
2. Confirm the robustness of retention time as an additional identification parameter
3. The importance of Mass Spectrometry (MS) settings, and
4. The impact of the bioinformatic treatment of MS data then MS/MS fragmentation data for dipeptide identification.

Overall, this work provides a guideline for comprehensive analytical and computational framework for the study of food-derived dipeptides, contributing to a better characterization of their diversity.