New DNA sequencing techniques are increasing the amount of data that can contribute to improving the precision of species definitions and genome annotations. In the bioinformatics laboratory, we are developing and applying genome-scale tools and procedures to predict microbial behavior and select microbial strains or enzymes with outstanding capabilities. Our studies apply to strains of agronomic, clinical, or industrial interest. Additionally, we work with metagenomics and machine learning to study complex environments and make biological predictions. We focus on the following topics:

Gene Mining and Microbial Profiles
The biological functions encoded in an organism eventually determine its phenotype and behavior in the environment. However, predicting the function of a protein is a complex task, with current in silico solutions centered on orthology. Recently, we developed a tool that establishes the bacterial functional profile based on the biological functions encoded in its genome. We mainly focus on plant growth-promoting rhizobacteria of the genus Bacillus.

Data Mining and Industrial Enzymes
There is a growing demand for new proteins or those with exceptional characteristics for application in various industrial processes and/or the development of genetically modified organisms. While the advent of genomics and data mining allows the search for hypothetical proteins homologous to known proteins or enzymes, many proteins available in databases share little identity with all studied proteins. In the laboratory, we use standardized procedures employing bioinformatics tools and statistical models to search databases for proteins with specific industrial purposes.

Taxonomic Resolution of Closely Related Strains
Despite the existence of guidelines and recommendations to ensure stability, reproducibility, and consistency in taxonomy, the methodology to circumscribe strains into species is still subjective and arbitrary. Additionally, the classification of species of many Firmicutes (Enterococcus, Lactococcus, Lactobacillus, or Bacillus) is not defined by a theoretical concept but generally by a pragmatic or industrial practice requirement. However, precise species assignment is vital for commercial acceptance for agronomic or industrial uses. Therefore, in the laboratory, we focus on assigning species to closely related phylogenetic strains, achieving consistency between phylogenetic, genomic, and phenotypic characteristics.

Metagenomics and Machine Learning
Metagenomics is the study of genetic material recovered directly from environmental samples, allowing the analysis of the diversity and functions of complex microbial communities without the need for cultivation. In our laboratory, we use metagenomics to study complex environments such as productive soil, where one gram of soil can contain more than 20,000 species. We apply machine learning, a branch of artificial intelligence focused on developing algorithms that enable computers to learn from data, understand biological complexity, and make predictions regarding soil yields and health.