The research record
Selected publications across spatial biology, applied AI and biological networks. For my complete publication record, visit Google Scholar.
Full Google Scholar profileNature Communications
Following tumour cells and their microenvironment through treatment, residual disease and recurrence.
A collaborative study combining single-cell RNA sequencing, spatial transcriptomics and imaging mass cytometry to investigate how treatment-tolerant cells persist in breast cancer.
ECML PKDD · Applied Data Science Track
Bringing pathologists into the learning loop to make spatial annotation more efficient.
Active learning uses tissue morphology and optional gene expression to select informative spots for expert annotation. The work evaluates this approach across breast cancer, colorectal cancer and kidney samples.
Communications Biology
Finding spatially distinct cellular niches without an external single-cell reference.
Chrysalis combines spatially variable gene detection with archetypal analysis to identify tissue compartments and their gene expression signatures. Its Python implementation integrates with Scanpy-based workflows.
Nature Methods
Learning protein representations that account for the cell types in which proteins act.
PINNACLE is a geometric deep learning approach to context-aware protein representations. It connects protein interaction networks with single-cell information to study protein function and therapeutic targets.
npj Precision Oncology
Resolving molecular subtypes in their tissue context, beyond a single bulk tumour label.
Spatial transcriptomics and single-cell RNA sequencing reveal how molecular programmes relate to tumour, stromal and non-neoplastic regions. Analysis scripts, notebooks and processed data accompany the study.
Molecular Systems Biology
Connecting communication between cells to the molecular signalling within them.
The OmniPath resource integrates molecular interactions, protein annotations and communication roles. R and Python interfaces make this prior knowledge accessible for multicellular omics analysis.
Bioinformatics
Exploring connected layers of biological knowledge to prioritise disease-associated genes.
An extension of random walk with restart to multiplex and heterogeneous networks, with an open-source R package for integrating different interaction types and gene–disease associations. The article appeared online in 2018 and in the 2019 journal issue.
Try another keyword or research area.
A curated selection, rather than an exhaustive bibliography. Find all publications on Google Scholar ↗