Head of Research Group: Dr. Balázs LIGETI
Members of the Group: Bendegúz FILYÓ, Dr. János JUHÁSZ, Judit JUHÁSZ, Dániel KRIZSÁN, Mátyás OSVÁTH, Dr. Csaba PONGOR
Contact: ligeti.balazs@itk.ppke.hu
Our research group focuses on neural-network- and sequence-based representations that capture large-scale genomic and evolutionary context. A key question in quantitative bi-ology is how to uncover novel patterns and structures in biological data, which is crucial for modeling, predicting, and manipulating complex biological systems like a microbiome . Our most recent research focuses on understanding the complex relationships character-izing the microbiome, such as phage-bacteria interactions. Phages, which are viruses that infect bacteria, can influence the structure of the microbiome and may also serve as both therapeutics and biomarkers.
We designed and implemented a genomic language model, ProkBERT (Ligeti et al ., 2024), to solve such bioinformatics tasks. ProkBERT provides a reusable, neural-network-based representation that can be applied to classification, re-gression, or clustering tasks related to microbiomes. This allows us to efficiently address problems such as identifying plasmids or viruses in complex microbiomes, or identifying potential candidates for phage therapy.
ProkBERT operates directly on genomic data and was trained on large corpora of microbial sequence data (bacteria, viruses, archaea, and fungi). It allows transfer learning by providing reusable sequence representations. The model is ideal for solving classification, clustering, and regression problems.
Future research directions, collaboration opportunities
The Neural Bioinformatics Research Group offers innovative, AI-compatible bioinformatics solutions: we use artificial intelligence-based tools to design phages or plasmids, as well as to solve other genomic problems. We are open to collaborations in research, industry, and grant-based initiatives.
Key publications
- Juhász, J., Ligeti-Nagy, N., Bodnár, B., Juhász, J., Pongor, S., Ligeti, B. (2025). ProkBERT Pha-Style: accurate phage lifestyle prediction with pretrained genomic language models. BIOINFORMATICS ADVANCES
- Ligeti, B., Szepesi-Nagy, I., Bodnár, B., Ligeti-Nagy, N., & Juhász, J. (2024). ProkBERT family: genomic language models for microbiome applications. FRONTIERS IN MICROBIOLOGY, 14, 1331233.
- Juhász, J., Ligeti, B., Gajdács, M., Makra, N., Ostorházi, E., Farkas, F. B., ... & Szabó, D. (2021). Colonization dynamics of multidrug-resistant Klebsiella pneumoniae are dictated by mi-crobiota-cluster group behavior over individual antibiotic susceptibility: a metataxonom-ic analysis. ANTIBIOTICS, 10(3), 268.
