Oceanic and intestinal flora reveals the mystery of microbial genes

An international team generated an automatic learning algorithm to predict unknown genetic functions of microbes. The system examines and compares the large existing data of human and environmental microbiome metagenomes.

Knowing the genes of bacteria that are part of the human microbiome – the group of microbes that live in us – is important because these genes can explain the mechanisms of bacterial infection or the cohabitation with the host, resistance to antibiotics or, more generally, the many positive or negative influences that the microbiome has on human health.

Surprising as it may seem, we do not know much about the functions of microbe genes [19659005] Surprising as it may seem, we do not know much about the functions of the genes of microbes. This knowledge gap may be considered a "genomic dark matter" of microbes, and neither comparative genomics nor current laboratory techniques have been able to unravel it.

This challenge was solved by an international scientific collaboration between the Institute of Biomedical Research (IRB Barcelona) and two other research institutes, the IJS in Ljubljana (Slovenia) and the RBI in Zagreb (Croatia ). Their results were published in Microbiome .

Fran Supek, first computational biologist at the Genome Data Science Laboratory of the IRB of Barcelona, ​​who led the work, and the first author of the publication, Vedrana Vidulin, developed a new Computational methodology able to examine thousands of metagenomes at a time and identify the evolutionary signal that can predict the function of many genes from microbes.

This method, which badyzes the large data of human microbiomes (from the intestine and the skin, for example) and other metagenomes (soil or ocean, for example) , based on a particular type of automatic learning algorithm: it can generate "decision trees" to predict hundreds of different genetic functions by finding links between genes and at the same time predicting functions They develop in the microbial cell.

"This makes the algorithm not confused with noise in metagenomic data, which Gnifica is very accurate and can reliably provide a biological role for a large number of genes of unknown function." is interesting to note that it also offers many additional functions for already known genes, "stresses Supek.

The most important result of this research is the badysis of human microbiomes and d & # 39; Other metagenomic data, such as those from the ocean or from the ocean to the soil, allows researchers to badign hundreds of genetic functions that were beyond the scope of current methods of computational genomics. 19659004] "In other words, metagenomes allow scientists to see what genomes can not show them," says Croatian researcher, recently rewarded with a grant from the European Council. Research (ERC)

Diversity is Key

Researchers have found that different types of environments can predict different types of genetic functions. For example, ocean metagenomes can be used to predict which genes the bacteria use for photosynthesis, whereas this could not be discovered from the bacteria that inhabit the human intestine. On the other hand, the intestinal microbiome has been very useful in predicting important genes for the pathogenesis, the metabolism of alcohol or the biosynthesis of certain amino acids, whereas these functions would have been difficult to discover. studying the microbiomes of the environment.

] The authors conclude that, through the application of machine learning, a large and diverse set of environments allows us to experience many different gene functions in microbes. "Computational methods like this shed light on" dark matter "(the large number of genes in bacteria and archaea whose function is not yet understood) in microbial genomes," Supek says. .

be validated experimentally, and once this happens, new relevant genes could be discovered to explain how bacteria shape the ecosystems around us and also our internal ecosystem, the human microbiome.