Scientists at the Luxembourg Centre for Systems Biomedicine (LCSB) at the University of Luxembourg have developed a new bioinformatics tool that enables them to identify pathogens faster and more accurately than with the methods currently used in diagnostics.
The team led by Professor Paul Wilmes, head of the LCSB's Systems Ecology group, uses metagenome data: all genome snippets from samples that may contain pathogenic organisms are sequenced in high-throughput. PathoFact compares the gene sequences with its own database. In this way, the bioinformatics tool identifies genes of microbes that are either important for their pathogenic effect or that make the bacteria resistant to antibiotics. This allows researchers to determine which pathogens are responsible for an infection and - in future clinical applications - to suggest suitable therapies. PathoFact also helps scientists to better understand the influence of microbes on the development of chronic diseases such as Parkinson's or diabetes.
Determining the origin of infections is still done using methods similar to those used 120 years ago in the days of Robert Koch and Louis Pasteur: bacteria, for example, are isolated from patient samples, cultivated and then determined. It often takes several days before one knows exactly which infection the patient is suffering from and how to treat it. "In the age of high-throughput genome sequencing, this should be faster," says Paul Wilmes: "To do this, you have to further develop the right bioinformatic techniques and combine them appropriately."
Among other things, PathoFact can use real-time data derived from metagenome sequencing. "This involves sequencing all the genes of all the microorganisms that are in a sample," explains first author Laura de Nies: "While the sequencing is still running, PathoFact can compare the information obtained with its own gene database." There, it searches for genes that are known to be responsible for virulence factors or antibiotic resistance. Virulence factors can be proteins that ensure the survival of bacteria in the human body. Or toxic metabolites that make the body sick.
Scientists already know the gene sequences for many virulence factors and microbial structures that are responsible for antibiotic resistance. They are stored in the database. Others, however, are completely new. "The proteins for which the genes code are similar to already known structures. They have certain features that are characteristic of virulence factors or antimicrobial resistance," says de Nies. Now the researchers can gain fundamentally new knowledge about pathogens and identify species that were previously unknown for their pathogenic effect.
Chronic diseases such as diabetes or Parkinson's disease have been at the centre of the investigations so far: in the case of such diseases, there is the now quite well-documented assumption that changes in the composition of the microbial communities - the microbiome - are involved in the development of the diseases. "The bacteria that live on and in us are in constant competition with each other," says Prof. Wilmes: "Certain living conditions can lead to harmful organisms multiplying more and triggering a disease with their toxic metabolic products." With the help of PathoFact, scientists can now detect such shifts in the microbial ecosystem faster and more accurately than before - and advance their basic research into chronic diseases.
Paul Wilmes also wants his development to benefit medical diagnostics: It should enable doctors to better predict severe courses of COVID-19 infections. This viral disease is often accompanied by so-called co-infections. The causative agents are bacteria or viruses other than SARS-CoV-2. "We can identify these pathogens more quickly than before. This opens up better treatment options for medical professionals to prevent severe COVID-19 courses in the future," says Wilmes: "In order for PathoFact to be used in clinical diagnostics, we plan to work together with a large Luxembourg diagnostics company. We will analyse samples that the company examines according to established procedures in parallel. In this way, we want to further develop and validate the high accuracy of our method. This should make it possible to treat dangerous microbial infections faster and more specifically in the future."
References: de Nies, L., Lopes, S., Busi, S.B. et al. PathoFact: a pipeline for the prediction of virulence factors and antimicrobial resistance genes in metagenomic data. Microbiome 9, 49 (2021).