Antibiotic resistance is making it increasingly difficult to treat bacterial infections. Phages, which attack bacteria, are presented as an alternative to normal antibiotic treatment. However, identifying which phage is effective against each bacterium is complex. This study, led by Robby Concha-Eloko, Beatriz Beamud, Pilar Domingo-Calap and Rafael Sanjuán , proposes the use of artificial intelligence to facilitate this prediction process.
For the development of the model they have used the bacterium Klebsiella, included in the WHO priority list of bacterial pathogens, responsible for serious hospital infections and with a high resistance to antibiotics. The Klebsiella bacteria are protected by polysaccharide capsules that prevent antibiotic activity, as well as phage entry. To overcome this barrier, many phages produce depolymerases, enzymes that break down these capsules and allow the bacteriophage to enter to infect the bacteria and contribute to their treatment.
However, the enormous genetic diversity of these capsules -more than 100 serotypes of these structures have been recorded in Klebsiella- has made it difficult to predict which phage may be the right one to get through the capsule and infect the bacteria. In turn, this great variety of capsule serotypes makes Klebsiella an ideal model for studying the interaction between phages and capsules.
The research team has developed a pioneering tool that leverages genetic information from thousands of Klebsiella bacteria and their "dormant" (pro-phage) viruses integrated into their genome. By analysing more than 74,000 programs and almost 20,000 depolymerase sequences, the research staff has created a database which associates each enzyme with the type of bacterial capsule that can degrade
To this end, the research team has developed a pioneering tool that harnesses the genetic information of thousands of Klebsiella bacteria and their "dormant" (pro-phage) viruses integrated into their genome. By analysing more than 74,000 programs and nearly 20,000 depolymerase sequences, the research staff has created a database that associates each enzyme with the type of bacterial capsule it can degrade.
Using advanced machine learning techniques and artificial intelligence models inspired by natural language processing, similar to those used by automatic translators, they have succeeded in predicting with great precision the "tropism" or specificity of each depolymerase, that is, what types of bacterial capsule can recognize and destroy.
This study provides a key breakthrough for phage-based biotechnology or its components, as it makes it possible to predict their specificity. This is essential for designing future applications, such as solutions against biofilm, the protective structure formed by some bacteria to adhere to surfaces and resist treatments.
Biofilms are increasingly recognized as a major obstacle in the treatment of infections. In fact, they have been shown to be involved in the chronicity of diseases such as cystic fibrosis, chronic injuries, prosthesis-related infections or urinary tract infections.
"The use of depolymerase, either in combination with current treatments (antibiotics or antimicrobial peptides) or potentially as an immune system enhancer, can address problems related to biofilm production, leading to a decrease in the risk of treatment failure", explains Robby-Concha.
"Unlike the traditional method, which relies on a tedious process of searching and testing for phages to find an effective depolymerase, artificial intelligence models allow us to predict its specificity in silico", Robby-Concha, researcher at I2SysBio
"Unlike the traditional method, which relies on a tedious process of searching and testing for phages to find an effective depolymerase, artificial intelligence models allow us to predict its specificity in silico," says the researcher. In this sense, the method demonstrated in the study allows the generation of depolymerase libraries that can be used to extract the most effective enzyme, optimizing the degradation of the capsule and subsequently the biofilm.
According to Robby-Concha, one of the developers of the pioneering tool, although Klebsiella has been used as a model, this methodology can be used against any other capsule-producing bacteria. This includes most of the priority pathogens collected by WHO.
In summary, this study proposes the resolution of the predictions of phage-host interaction in two ways. Firstly, by exploiting the contained bacterial genomic data (profagos), which allows important training data to be obtained; and secondly, by proposing an architecture that allows the model to be trained with all bacterial species at once (in an integrative way).
Don’t miss the interview with Robby Concha in our EuroPark section
Source: UV News