Annual Conference 2025 of the Association for General and Applied Microbiology - Jahrestagung 2025 der Vereinigung für Allgemeine und Angewandte Mikrobiologie

BacDive: The core database for prokaryotic strain data

Computational microbiology

Since its initial release more than a decade ago, the bacterial diversity database BacDive (https://bacdive.dsmz.de) has grown tremendously in terms of content and functionality. Today, BacDive is a comprehensive resource covering the phenotypic diversity of prokaryotes with data on taxonomy, morphology, physiology, cultivation, and more. As the leading database for strain-level information on bacteria and archaea, it has been designated as an ELIXIR Core Data Resource and a Global Core Biodata Resource. These distinctions underscore its importance to the wider microbiological research community. As part of the newly established DSMZ Digital Diversity infrastructure, BacDive is now closely linked to several other databases of fundamental importance to the life sciences, including BRENDA, LPSN and SILVA. These databases will be further developed in a coordinated manner and benefit from frequent data exchange.Here we present the current status and recent developments in BacDive. The database currently contains 2.6 million data points for 97,334 strains. This includes the world's largest collection of Analytical Profile Index (API) test results, which have now been fully translated into easily understandable and searchable data fields. Increasingly, the standardized high-quality data provided by BacDive is being used to train artificial intelligence models. The high-confidence genome-based predictions produced by our own models are now also being used to fill content gaps in the database. We further present a novel BacDive knowledge graph that provides powerful new search capabilities via a SPARQL endpoint to directly search and analyze the knowledge provided through BacDive in a standardized way, supported by a new descriptive ontology.

Poster

Prof. Dr.

EDGAR 3.5: Advancing comparative genomics with new tools for microbial genome analysis

IntroductionThe significant increase of whole genome sequence data availability, made possible by next generation sequencing methods, has caused the emergence of a variety of tools for the comparative analysis of related genomes. The web-based, EDGAR platform, is one of the most established tools, to highlight similarities and differences, conserved and divergent regions within microbial sequence data. EDGAR is being constantly updated for the last 15 years to serve the emerging needs of the scientific community. This work focuses on the latest developments on the EDGAR platform and expands on upcoming changes.ObjectiveSince its inception, the average number of genomes analyzed per EDGAR project has steadily increased, as has the variety of methods employed by its community. As a consequence, the development of EDGAR pursues two main goals: First, to create the necessary infrastructure to accommodate the mounting data volumes. Second, to constantly improve the EDGAR website with new ways for scientists to examine their genomes.MethodsWe iterated further on our recently released functional category features and will soon offer the interactive exploration of KEGG subcategories organized in a sunburst plot. The GO term categorization has received a major upgrade, with it now being possible to calculate an enrichment analysis between a chosen subset of genomes. This is being additionally expanded upon in the form of a GO-Net visualization, which will make it possible to browse through the web of correlated genes and hierarchical GO-terms. Furthermore, we are working on a feature that enables our users to find, save, visualize and compare conserved syntenic gene clusters. The core ortholog calculation has been rewritten, making it faster by a factor of 50x, enabling EDGAR to scale into the ultra-high-throughput era. This will allow us to use binned subsets of the BakRep project to provide a significantly scaled-up version of EDGAR's public database in the near future.ResultsEDGAR continues to grow its capabilities, offering an increasingly diverse array of analysis methods for comparative genomics and phylogenomics in a single software framework. Our infrastructure and web platform offer researchers a time-effective way to gain various perspectives on their microbial genomes. For unpublished data, EDGAR offers private, password protected projects. Our web server is accessible at: https://edgar3.computational.bio

Prof.

NFDI4Microbiota – Enhancing microbiota research through data integration and digital transformation

Over recent years, advances in omics data generation have opened up significant opportunities for researchers working with "big data." As a result, the challenges in the field have evolved, shifting from data collection to the reuse, interpretation, and analysis of vast, heterogeneous datasets. The NFDI4Microbiota consortium was formed to meet these new demands by offering data access, analytical services, training, and infrastructure to support the broader microbiological community.NFDI4Microbiota aims to serve as a central hub established to assist both the German and international microbiological research communities. It comprises ten partner institutions backed by five professional societies and over 50 participating organizations. The consortium aims to accelerate the digital transformation within the microbiological community by providing solutions for data management through cutting-edge computational methods. It also seeks to streamline the entire research workflow, from initial data generation to the final stages of publication, including data submission. The consortium offers specialized training programs, infrastructure, and advanced computational tools to achieve this streamlining process.The wide array of tools, methodologies, and resources offered by NFDI4Microbiota is designed to facilitate the transformation of new and existing data into meaningful scientific knowledge. Training opportunities are available in open science, research data management, databases, programming, data science, and data analysis. Furthermore, implementing a cloud-based infrastructure provides access to analytical tools, a knowledge base, and systems for analyzing, integrating, and storing microbiological data.By creating this centralized resource, NFDI4Microbiota aims to become the essential connection for microbiota research in Germany and beyond. It offers researchers the infrastructure, expertise, and easy access to tools and training that enhance their work while supporting Open Science principles and ensuring that data remains Findable, Accessible, Interoperable, and Reusable (FAIR) in the future.

Computational modelling of metabolic interactions in a virtual colon

Question Metabolic interactions are important for the function, health, and disease of host-microbial systems. Reconstructing metabolic models of organisms from host and microbial genomes allows us to study these interactions. However, these interactions are often considered static, neglecting spatial and temporal factors. MethodsThe software BacArena enables the co-simulation of those models in a computational space over time and has been widely used to decipher such phenomena. Here, we present Virtual Colon, an application of this software which is specialised to replicate the structure of the mammalian colon. The host models are separated from the bacterial ones, which reside in a simulated lumen including mucus layers.ResultsOur proof-of-concept experiments include transcriptomics-based, context-specific colonic models combined with individual bacterial species and a bacterial collection, such as the extended simplified human intestinal microbiota (SIHUMIx). The Virtual Colon notably visualises those interactions either predicting known ones (e.g., short-chain fatty acids exchange) or novel ones (e.g., amino acids exchange). ConclusionsVirtual colon enhances genome-scale metabolic modelling, providing new insights into spatiotemporal host-microbiome systems. Like BacArena, it is freely available and can be adapted for other spatially structured environments (www.github.com/maringos/VirtualColon, www.doi.org/10.1101/2024.06.11.598488).

­­Comprehensive hybrid genome assembly reveals mobile genetic elements, antimicrobial-resistance genes and antimicrobial peptides in Bacteroidota spp. using machine learning approaches

Bacteroidota species, dominant members of the human gut microbiome, are emerging as promising candidates for Next Generation Probiotics (NGPs) due to their extensive metabolic capabilities, antimicrobial properties, and beneficial interactions with the host immune system. The emergence of Bacteroidota spp. as NGPs necessitates detailed genomic characterization to understand their therapeutic potential. This study presents a comprehensive genomic and pangenome analysis of 18 Bacteroidota isolates from diverse sources including mouse cecum, animal feces, and human stool samples, utilizing an innovative hybrid sequencing approach combining Illumina short reads and Oxford Nanopore long reads. Our optimized hybrid assembly pipeline, employing Unicycler with Canu-corrected long reads, achieved exceptional genome completeness (≥99.68%) with minimal contamination. Three isolates achieved single-contig circular genomes, while seven others demonstrated minimal complexity in their assembly graphs. The assembled genomes showed consistent coding density (90.5-98.5%) with gene counts ranging from 3,968 to 5,150, as annotated by Bakta. Taxonomic classification using GTDB-TK identified the isolates as members of the Bacteroides and Phocaeicola genera, with Fast-ANI scores above 97.07%. Comprehensive screening using multiple databases (NCBI, CARD, and Resfinder) through ABRicate revealed the presence of diverse antimicrobial resistance genes among isolates. Novel potential antimicrobial peptides were predicted using an enhanced prediction approach combining hyperparameter tuning of machine learning models, optimized amPEPpy v1.1.0, trained on ADAPTABLE and AMPSphere databases, complemented by Macrel v1.4.0 for high-quality AMP candidate identification and small open reading frame prediction. Characterizing the isolates experimentally demonstrated significant probiotic potential, with all strains showing capacity to inhibit biofilm formation against clinically significant pathogens. Notably, Bacteroides caecimuris and Bacteroides muris exhibited exceptional biofilm inhibition and quorum quenching activity. This comprehensive genomic, pangenomic and functional characterization establishes these Bacteroidota strains as promising candidates for therapeutic applications, advancing our understanding of their potential as next-generation probiotics.

Computational microbiology 1

Termin

Programm