attempto online


Only 3% of potential bacterial drug sources known

Bioinformatics study pinpoints the most promising bacterial candidates for successful future investigations

How much do we know about natural resources for drugs? A bioinformatics study indicates that we might have so far uncovered only 3% of potential bacterial drug sources, pinpoints the most promising bacterial candidates, and therefore could help decide, which bacteria deserve further attention for successful future investigations. This is good news at a time when new drugs are desperately needed due to the emerging antibiotic resistance crisis.

Where to search for new drugs?

Natural Products of bacterial origin have been studied for decades now as sources of drugs, such as antibiotics. However, the rate of new drug discovery has stagnated. The question raised is:

Is there more to discover in nature or has this source of drugs run out?

If there is more out there, where to search for it? Which bacterial groups look promising and how many are they?

Genome mining, a bioinformatics strategy, can play a key role in identifying those promising groups and increase the probability of successfully discovering a new bacterial drug resource. With Genome mining it is possible to detect the genes that are responsible for the production of a certain compound. The mining criteria are being set for a target-oriented search through the genome to identify the sequences of interest. Those sequences are usually genes.

Solid data for better decision-making

Starting point was the observation, that the decision for a certain genus at the very beginning of every search for a new compound is not trivial. Many labs go for well-known Streptomyces since it seems to be the most divers genus. But reliable data on which to base this decision do rarely exist. Researchers from the Cluster of Excellence CMFI at the University of Tübingen, together with colleagues from Wageningen University, Scripps Research Florida, Helmholtz Centre for Infection Research (HZI) and the German Center for Infection Research (DZIF) wanted to establish solid data and distinct criteria to be used for better decision making. In a first step collected bacterial genomes publicly available from databases and from microbiome studies.

With a genome mining method Biosynthetic Gene Clusters (BGCs) were identified. BGCs, that is, a bunch of genes that cluster together in the genome and encode everything necessary for the biosynthesis of the same Natural Product. These data gave insights in how many and what kind of compounds each microbe can synthesize.

However, it is possible that two completely different bacteria produce the exact same compound, either due to their common ancestor or due to the acquisition of the necessary genes via Horizontal Gene Transfer (HGT), that is, — in contrast to vertical gene transfer from parent to offspring — the transfer of genetic material between different organisms.

For that reason, looking only at the Gene Clusters was not enough. With tools from bioinformatics the team grouped them into so-called Gene Cluster Families according to their similarity. The results showed that a genome could include five Gene Clusters that span over four Gene Cluster Families, or it could include fifteen Gene Clusters that span over three Gene Cluster Families.

Having quantified the biosynthetic diversity of known bacteria, the researchers were able to build a computational model to predict how much diversity could be found in the bacterial kingdom.

”The result is a very hopeful message for the future of Natural Product research. Especially since the 3% is the highest possible number - we could still be underestimating the bacterial kingdom's hidden potential.” says Athina Gavriilidou, first author of the study.

Select the most promising candidates

Knowing that there is more to discover, the research team still needed to point out which bacteria are really worth their time and effort to be studied further.

“Quite often there is only rediscovery of known compounds at the end of time-consuming work in the laboratory. Therefore, we do not only need to identify bacteria that produce much, but also those who produce a diversity of compounds. This way a research strategy becomes more effective and the chances of finding something new increases,” says Prof. Dr. Nadine Ziemert, Principal Investigator and Board member at the Cluster of Excellence CMFI and head of the study.

Instead of looking at the bacterial kingdom as a whole, the researchers inspected each bacterial genus (group of closely related species). By doing so, they were able to create a ranking based on the predicted biosynthetic diversity of all discovered genera from most to least promising. Based on the biosynthetic diversity of the genera, uniform taxa could be created. Those equivalences enable for a more precise comparison of the genera. Streptomyces, for example, was subdivided into 21 smaller genera by applying the new metrics. This puts its preeminence in the search for new agents into perspective.

The results were combined with the so-called phylogenetic tree of bacteria, which depicts evolutionary relationships. An exciting yet complicated image emerged, that allows to visualize the biosynthetic capacity of all identified bacteria.

“Our results show how important bioinformatics are, when we think of interdisciplinary research in the life sciences and effective ways to search for new drugs. Our findings enable researchers to focus on the most promising spots in the bacterial kingdom. It is nice to see how much ungrazed ground we have in front of us. Having these promising fields in sight is an exciting development while facing a growing number of drug resistances and less newly discovered compounds. That is our next challenge: We have the genes now. What we need to find are the new actual compounds,” Ziemert says.

Leon Kokkoliadis


Gavriilidou A, Kautsar SA, Zaburannyi N, Krug D, Müller R, Medema MH, Ziemert N. Compendium of specialized metabolite biosynthetic diversity encoded in bacterial genomes Nat Microbiol. 7(5). (2022) doi: https://doi.org/10.1038/s41564-022-01110-2.

Scientific contact:

Prof. Dr. Nadine Ziemert
University of Tübingen
Cluster of Excellence “Controlling Microbes to Fight Infections” (CMFI) Interfaculty Institute of Microbiology and Infection Medicine Tübingen Translational Genome Mining for Natural Products
 +49 7071 29-78841
nadine.ziemertspam prevention@uni-tuebingen.de 


Leon Kokkoliadis
Cluster of Excellence “Controlling Microbes to Fight Infections” (CMFI)
 +49 7071 29-74707
Leon.kokkoliadisspam prevention@uni-tuebingen.de