Dr. Gregorio Iraola is a computational microbiologist. His post-graduation studies started with a Master in Bioinformatics before obtaining a PhD in Biology focused on microbial genomics at the Universidad de la República and the Institut Pasteur Montevideo in Uruguay. Since 2017 he is a staff Associate Researcher at the Bioinformatics Unit in the Institut Pasteur Montevideo. Also, since 2015 he has been a visitor scientist at the Wellcome Trust Sanger Institute and the Institut Pasteur Paris. From Uruguay, he leads several research lines aiming to develop and apply computational approaches for studying the microbial world. His work has been focused on understanding the evolution of viruses that affect livestock and pets using phylodynamics; and fundamentally on uncovering the evolutionary forces shaping the genomes of zoonotic bacteria like Campylobacter, Leptospira and Mycobacterium. Among his ongoing projects stands out a joint Latin American effort to study the population dynamics of Clostridium difficile using genomic epidemiology approaches. Recently, he became interested in complementing his work in pathogenomics with microbiome approaches, specifically by applying cityscale metagenomics to analyze antibiotic resistance dynamics in enterobacteria from urban environments. More recently, he got involved in science communication as a columnist aiming to bring microbiology and genomics closer to the society.
Bacterial pan-genome estimation takes a walk inside taxonomy
The bacterial pan-genome concept was coined in 2005 to describe the union of genes shared by a certain set of genomes. This notion was then formalized by the distributed genome hypothesis, which states that the gene pool in a bacterial population is more complex and larger than that found in the genome of any single strain. Since then, the pan-genome approach has been extensively
applied to characterize the genetic variability of bacteria at the population level (i.e. many genomes of different strains from the same species). Accordingly, methodological approaches for estimating bacterial pan-genomes have been developed to deal with the genetic variability observed between members of the same species. However, some relevant questions about the biology of bacteria
require to be answered at deeper taxonomic levels, for example which gene clusters differentiate pathogenic from microbiota-associated strains within the genus Finegoldia1, which is the phylogenetic structure and taxonomy of the order Chlamydiales2 or which are niche-specific marker genes within the pan-genome of the class Epsilonproteobacteria3. Dealing with deeper taxonomic
units suppose to work with more divergent microorganisms, hence with evolutionary distant genes whose orthology relationships are more elusive to be easily determined. For this reason, we developed Pewit (https://github.com/iferres/Pewit), a software tool that efficiently estimates bacterial pan-genomes at deeper taxonomic ranks using a combination of sequence identity and protein architecture similarity metrics. We tested Pewit using both simulated pan-genomes and real datasets covering a representative set of bacteria, demonstrating that it outperforms state-of-the-art softwares for pan-genome estimation. Recently, we used Pewit to identify a novel group of species within the genus Leptospira with a distinctive accessory genome from pathogenic species that
correlates with their reduced pathogenic potential4.