Research

CEMERAI is designed around four interconnected scientific themes that connect data generation, modelling, prediction, and validation.

Research strategy

The centre integrates glycoproteomics, proteomics, lipidomics, structural biology, mass spectrometry, molecular dynamics, and AI to move from descriptive molecular catalogues toward predictive, multiscale membrane models.

CEMERAI overview figure

CEMERAI goals table

Strategic Goal	Key Scientific Questions	Core Experimental Platforms	AI/Computational Components	Expected Outputs / Impact
1. Multiscale modelling of mitochondria and red blood cells	How do lipids, proteins, and glycans organize to maintain cell membrane integrity?	glycoproteomics, lipidomics, crosslinking and structural MS, cryo-EM/-ET, MD simulations	Integrative AI for multimodal coordination, ML-based feature extraction from Cryo-ET; generative AI	Predictive atomistic-mesoscale RBC/mitochondrial cell membrane model
2. Modelling drug and disease-related membrane variations	How do cell membranes re-model with disease or stress or response to drugs? How do drugs enter membranes?	glycoproteomics, lipidomics, crosslinking and structural MS, cryo-EM/-ET, MD simulations, cell line and animal models	Predictive AI for dynamic cell-membrane remodelling; mapping altered lipid composition to biophysical profiles; causal modelling linking omics and phenotype	Mechanistic understanding of membrane changes in pathology; biomarkers for disease progression
3. Comparative and predictive modelling of diverse cell membranes	What are the conserved and divergent structural principles across cell types and mitochondria isolated from different cell types?	Single-cell lipidomics, glycoproteomics, native MS, Cell Atlas imaging	Integrative AI for cross-cell comparison; transfer learning between models	Cross-cellular membrane atlas; digital reference for drug response and tissue-specific therapeutics
4. Use of synthetic protein design for diagnostics and therapeutics	Development of advanced data-driven tools for diagnostics and treatment	Synthetic biology, high-throughput binder design	Generative AI for peptide/protein binder design	New therapeutic and diagnostic molecules; open-access AI models for the life science community

Themes

Multiscale modelling of erythrocytes and mitochondrial membranes

Combining glycoproteomics, lipidomics, proteomics, structural biology, structural mass spectrometry, and integrative AI models to create testable cell membrane models.

Leads:

Modelling membrane alterations in response to external stress, drugs, or disease

Identifying plasma membrane and mitochondrial remodelling events during pathological, drug-induced, or stress-induced states.

Leads:

Comparative and predictive modelling of diverse cell membranes

Building cross-cell-type models to capture conserved and specialised principles of membrane organisation for diagnostics and rational therapeutic design.

Leads:

Synthetic protein design for diagnostics and therapeutics

Using AI tools to develop synthetic binders against key membrane proteins, including glycoproteins with high specificity.

Leads:

Generation of predictive AI models to improve drug transport into cells

Here we will develop generative and predictive AI models to improve our understanding of drug transport into cells.

Leads: