View Proposal

---

Proposer

Marta Vallejo

Title

Explainable Machine Learning for Interpreting Tissue NMR Signals Using Histology and Proteomics (in collaboration with the University of Aberdeen)

Goal

The project aims to develop an explainable machine learning framework that integrates NMR signal profiles, histological images and proteomics data to identify how tissue structure and molecular composition contribute to changes in NMR measurements.

Description

Nuclear Magnetic Resonance (NMR) provides rich information about the physical and molecular properties of biological tissues, but understanding which tissue characteristics are responsible for changes in the measured signals remains a significant challenge. This project will investigate the relationship between NMR signal profiles acquired across different magnetic field strengths, histological images of the corresponding tissue samples, and proteomics data describing their molecular composition. The student will develop machine learning methods to analyse these complementary datasets, identify associations between tissue structure, protein expression and NMR signals, and build interpretable models that explain how specific biological changes influence the measured NMR profiles. The project offers experience in multimodal data analysis, explainable artificial intelligence, biomedical image analysis and computational biology, and is carried out in collaboration with the University of Aberdeen with the potential to contribute to a scientific publication.

Resources

(1) Paired NMR signal datasets acquired across multiple magnetic field strengths, (2) Histological images from the corresponding tissue samples, (3) Proteomics datasets matched to the tissue samples.

Background

Nuclear Magnetic Resonance (NMR) is widely used to investigate the physical and molecular properties of biological tissues. However, the biological factors responsible for changes in NMR signals are often difficult to interpret. Recent advances in machine learning provide new opportunities to integrate NMR measurements with histological imaging and proteomics, enabling researchers to relate signal variations to tissue structure and molecular composition. Such approaches have the potential to improve our understanding of tissue pathology and support the development of more informative and interpretable biomedical biomarkers.

Url

Difficulty Level

Variable

Ethical Approval

Full

Number Of Students

2

Supervisor

Marta Vallejo

Keywords

nmr, multimodal machine learning, explainable ai, histology, proteomics, biomedical signal analysis, computational pathology, tissue characterisation

Degrees

Bachelor of Science in Computer Science
Master of Engineering in Software Engineering
Master of Science in Artificial Intelligence
Master of Science in Computing (2 Years)
Master of Science in Data Science
Master of Science in Software Engineering
Bachelor of Science in Computing Science
Bachelor of Engineering in Robotics
Master of Science in Robotics with Industrial Application
Postgraduate Diploma in Artificial Intelligence