Roni, R. A., Bauby, P. C., Hernandez Soto, G., Liu, C., & Zhao, J. (2026). Pea protein gel hardness across published studies and laboratory validation: Effects of formulation and processing context.
← Back to all work
Food Protein Functionality & AI-Based Prediction
Graduate research combining experimental and literature-derived data with machine learning and deep learning to investigate and predict food protein gelation and functional behavior.
- Question
- How do formulation and processing conditions shape food protein gelation and related functional behavior, and how well can these properties be predicted from structured experimental and literature-derived data?
- Why it matters
- Gelation is one of the functionalities that decides whether a protein ingredient can replace a conventional one. Reported results are scattered across studies with inconsistent conditions, so comparable, structured data is needed both to understand gelation and to predict it.
- My role
- Graduate Research Assistant in the Zhao Food Chemistry Lab. First author of the pea-protein gel-hardness analysis and its PPIC 2026 poster; co-author on related collaborative machine-learning posters.
- Status
- Ongoing M.S. thesis research (expected May 2027). Pea-protein manuscript in preparation.
- Methods
- Extraction and curation of gelation data from published literature
- Laboratory validation of selected conditions
- Structured datasets linking protein concentration, pH, salts and additives, temperature, and method to functional outcomes
- Statistical analysis and Python-based machine-learning workflows
- Deep-learning approaches (in development)
- Outputs
- PPIC 2026 poster (first author)
- Pea-protein manuscript in preparation (first author)
- Three co-authored team posters (ACS Spring 2025, IFT FIRST 2025, Pacifichem 2025)
Overview
My M.S. thesis, “Prediction of Food Protein Gelation Properties Using Machine Learning and Deep Learning,” brings together experimental and literature-derived data to study and predict how food proteins form gels under different formulation and processing conditions.
Pea protein gel hardness across published studies
One component of this work is my first-author quantitative analysis of pea protein gel hardness reported across the published literature, organized by formulation and processing context and complemented by laboratory validation.
It was presented as “A quantitative analysis of pea protein gel hardness from published literature” at the Plant Protein Innovation Center (PPIC) 7th Annual Research Spotlight Meeting (2026), and is being written up as “Pea protein gel hardness across published studies and laboratory validation: Effects of formulation and processing context” (manuscript in preparation).
Collaborative machine-learning work
I am a co-author on related team posters that apply machine-learning, deep-learning, and AI models to protein gelation and gel-hardness prediction, presented at ACS Spring 2025, IFT FIRST 2025, and Pacifichem 2025.
Thesis direction
The thesis builds on these studies to predict food protein gelation properties more broadly. Structured datasets and machine-learning workflows are in active use; deep-learning approaches are still being developed. Results will be reported through the thesis and associated publications.
Related outputs
Poster · first author
Rabiul Alam Roni, Philippine Curail Bauby, Giovanni Hernandez Soto, Changqi Liu, & Jing Zhao. (2026). A quantitative analysis of pea protein gel hardness from published literature.
Co-authored poster
Meryem El Alaoui Hassani, Ali Raza, Harrison Ndiba, Rabiul Alam Roni, Dominik Grabinski, David A. Hecht, & Jing Zhao. (2025). Prediction of Protein Gelation Properties Using Artificial Intelligence Models.
Co-authored poster
Meryem El Alaoui Hassani, Harrison Ndiba, Ali Raza, Mozhgan Esmaeelian, Rabiul Alam Roni, David A. Hecht, Dominik Grabinski, & Jing Zhao. (2025). Predicting Protein Gel Hardness Using Machine Learning and Deep Learning Models.
Co-authored poster
Ali Raza, Meryem El Alaoui Hassani, Harrison Ndiba, David A. Hecht, Rabiul Alam Roni, Mozhgan Esmaeelian, Dominik Grabinski, & Jing Zhao. (2025). Predicting protein gelation properties through modeling using published literature results.