Personal Statement
Leo is an evolutionary computational biologist interested in (1) large-scale omics —phylogenomics, microbiomics, functional genomics— and (2) statistical modelling with uncertainty. His goal is to bridge statistical tools and evolutionary analyses and apply them to biomedical sciences and healthcare.

He is based at the Signal Processing Group and is part of the Data Action Accelerator of the Civic Health Innovation Labs (CHIL), with close links to the Microbiome Innovation Centre (MIC).

He did his PhD under the supervision of Hirohisa Kishino, followed by postdoctoral work in Spain, Switzerland, and at Imperial College London. His work focuses on statistical evolutionary biology, including genomic heterogeneity in the Tree of Life, the molecular clock hypothesis, and protein structure for evolutionary inference. He specialises in Bayesian phylogenetics, developing models and software for whole-genome settings and deploying them on high-performance computing systems. He has also applied statistical learning to phylogenomics and hyperspectral imaging.

Before joining the University of Liverpool, he was Head of Phylogenomics at the Quadram Institute (UK), where he worked on large-scale bacterial genomics. During this time, he contributed to the COVID-19 Genomics UK (COG-UK) consortium, conducting studies on SARS-CoV-2 genomes generated at the Institute. He developed scalable database search methods and fast phylodynamic models, and participated in other pandemic response initiatives, including the Real-time Assessment of Community Transmission (REACT) in the UK and efforts led by the Africa CDC.

Research Overview
Leo designs and implements scalable Bayesian models for genomes and diseases, for instance microbial genes of clinical relevance. He is also interested in engineering solutions for analysing and sharing One Health information, including biosurveillance.

Currently his main research involves (1) developing Bayesian phylogenomics for an integrated view of microbes together with their hosts and environment, and (2) extending evolutionary models for application in general Statistical Learning problems.