Research
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.
Phylogenetics and phylogenomics
Leo develops statistical and computational models to study how genes evolve within species and across the tree of life. He designs Bayesian hierarchical methods to understand how genetic variation changes along genomes [1,2], to simulate the evolution of new genomes under complex biological scenarios [3], and to investigate the early origins of eukaryotic life [4]. His work also applies phylogenetic approaches—methods that reconstruct evolutionary history—to address diverse biological questions, such as estimating when marine species originated [5], identifying signatures of natural selection in genes [6], and tracing the geographical spread and evolution of pathogens [7].
Phylogenetics focuses on how individual genes evolve, by modelling the changes observed at homologous sites in a sequence alignment to reconstruct a single evolutionary history. Phylogenomics extends this perspective to the scale of whole genomes: because different genes experience distinct evolutionary processes—such as varying mutation rates, recombination, or birth and death—their evolutionary histories may not coincide. Leo is interested in both levels: site-level evolution within genes, as well as gene-level evolution across genomes, capturing heterogeneity and the complex biological processes that shape genome-wide evolutionary patterns.
[1] https://doi.org/10.1093/sysbio/syu082
[2] https://doi.org/10.1007/s10463-009-0259-8
[3] https://doi.org/10.1093/sysbio/syv082
[4] https://doi.org/10.1093/gbe/evac119
[5] https://doi.org/10.1371/journal.pone.0010363
[6] https://doi.org/10.1186/1471-2148-10-52
[7] https://doi.org/10.1038/s41467-022-32096-4
Microbial Genomics
Leo develops computational and statistical approaches to analyse large genomic datasets, particularly those related to virulence and antimicrobial resistance (AMR). He has experience working with large AMR databases, including designing idealised benchmarking datasets [1] and analysing resistance genes in both microbial cultures [2] and metagenomic data [3].
His work also uses evolutionary principles to improve how microbial genomes are analysed. He has developed methods to identify new phylogenetic markers [4], critically assess existing markers [5], and accelerate searches across large genomic databases [6]. The latter approach has been used to enable the rapid analysis of large and clinically important viral outbreaks [7].
[1] https://doi.org/10.1038/s41597-022-01463-7
[2] https://doi.org/10.1038/s41522-020-00178-0
[3] https://doi.org/10.1099/mgen.0.001284
[4] https://doi.org/10.1093/nargab/lqac003
[5] https://doi.org/10.1093/nargab/lqz016
[6] https://doi.org/10.7717/peerj.16890
[7] https://doi.org/10.1016/S2214-109X(21)00434-4
Algorithmic Bioinformatics
Leo develops algorithmic and data-engineering methods for biology, with applications ranging from biomedical analysis to genomic surveillance. His work includes computational methods to measure distances between evolutionary trees [1] and between genomes [2], allowing complex genomic information to be transformed into features that can be analysed using conventional machine-learning techniques [3,4].
He has also contributed statistical approaches to fundamental biological questions, including studying the limits of homology detection [5] and modelling antibody–protein binding affinities [6]. In addition, he has developed hyperspectral imaging algorithms for biological analysis [7], which have been applied to investigate cellular responses to biomaterials [8].
[1] https://rdrr.io/cran/phangorn/man/treedist.html
[2] https://github.com/leomrtns/amburana
[3] https://github.com/leomrtns/genefam-dist
[4] https://anaconda.org/channels/bioconda/packages/super_distance
[5] https://doi.org/10.1093/sysbio/syu041
[6] https://doi.org/10.1093/molbev/msm079
[7] https://github.com/leomrtns/specimage
[8] https://doi.org/10.1021/acsnano.8b06998