Pages

Other

From Axioms over Graphs to Vectors, and Back Again: Evaluating the Properties of Graph-based Ontology Embeddings

Year: 2023 Venue: Proceedings of the 17th International Workshop on Neural-Symbolic Learning and Reasoning, La Certosa di Pontignano, Siena, Italy, July 3-5, 2023 Authors: Fernando Zhapa-Camacho, Robert Hoehndorf Abstract Several approaches have been developed that generate embeddings for Description Logic ontologies and use these embeddings in machine learning. One approach of generating ontologies embeddings is by first embedding the ontologies into a graph structure, i.e., introducing a set of nodes and edges for named entities and logical axioms, and then applying a graph embedding to

Fully Geometric Multi-hop Reasoning on Knowledge Graphs with Transitive Relations

Year: 2026 Venue: The Semantic Web -- ESWC 2026 Authors: Fernando Zhapa-Camacho, Robert Hoehndorf DOI: 10.1007/978-3-032-25156-5_14 Abstract Multi-hop logical reasoning on knowledge graphs requires faithfully mapping the logical semantics to latent space. Current geometric embedding methods show to be useful on this task by mapping entities to geometric regions and logical operations to latent transformations. While a geometric embedding can provide a direct interpretability framework for query answering, current methods have only leveraged the geometric construction of entities, failing to

Genomic diversity and antimicrobial resistance of Staphylococcus aureus in Saudi Arabia: a nationwide study using whole-genome sequencing

Year: 2025 Venue: Microbial Genomics Authors: Mohammed S. Alarawi, Musaad Altammami, Mohammed Abutarboush, Maxat Kulmanov, Dalal M. Alkuraithy, Senay Kafkas, Robert Radley, Marwa Abdelhakim, Hind Aldakhil, Reema A. Bawazeer, Mohammed A. Alolayan, Basel M. Alnafjan, Abdulaziz A. Huraysi, Amani Almaabadi, Bandar A. Suliman, Areej G. Aljohani, Hassan A. Hemeg, Mohammed S. Almogbel, Meshari Alazmi, Abdulrahman S. Bazaid, Turki S. Abujamel, Anwar M. Hashem, Ibrahim A. Al-Zahrani, Mohammed S. Abdoh, Haya I. Hobani, Rakan F. Felemban, Wafaa A. Alhazmi, Pei-Ying Hong, Majed F. Alghoribi, Sameera

Genomic landscape in Saudi patients with hepatocellular carcinoma using whole-genome sequencing: a pilot study

Year: 2023 Venue: Frontiers in Gastroenterology Authors: Mazen Hassanain, Yang Liu, Weam Hussain, Albandri Binowayn, Duna Barakeh, Ebtehal Alsolme, Faisal AlSaif, Ghaida Almasaad, Mohammed AlSwayyed, Maram Alaqel, Rana Aljunidel, Sherin Abdelrahman, Charlotte A. E. Hauser, Saleh Alqahtani, Robert Hoehndorf, Malak Abedalthagafi DOI: 10.3389/fgstr.2023.1205415 Abstract Our findings indicate that most of the HCC patients possess cancer-related genetic variants, and the altered pathways in these patients exhibit similarities. Notably, resistant patients exhibit a higher frequency of aberrations in

Genomic landscape of retinoblastoma: Insights into risk stratification and precision pediatric Neuro-Oncology

Year: 2025 Venue: Neuro-Oncology Pediatrics Authors: Azza Maktabi, Yang Liu, Saleh Almesfer, Marwa Abdelhakim, Hind Aldakhil, Maxat Kulmanov, Deepak P Edward, Robert Hoehndorf, Malak Abedalthagafi DOI: 10.1093/neuped/wuaf017 Abstract Abstract Background Retinoblastoma is the most common intraocular malignancy of childhood, yet its genomic landscape remains incompletely defined, particularly in understudied populations. Beyond RB1 loss, the contribution of additional somatic and germline alterations to disease heterogeneity and clinical behavior is unclear. Methods We performed whole-exome

Genomics

Our genomics work develops resources and computational methods for population-scale genome analysis, with a particular focus on populations that have historically been under-represented in reference databases. Within the BORG group at KAUST's Computer Science Program, we build reference genome assemblies, pangenome graphs for the Saudi and wider Middle Eastern population, variant-calling and structural-variant pipelines, and analyses of antimicrobial resistance from whole-genome sequencing. Our distinctive angle is the tight integration of formal phenotype and function knowledge with sequence

GFVO: the Genomic Feature and Variation Ontology

Year: 2015 Venue: PeerJ Authors: Joachim Baran, Bibi Sehnaaz Begum Durgahee, Karen Eilbeck, Erick Antezana, Robert Hoehndorf, Michel Dumontier DOI: 10.7717/peerj.933 Abstract Falling costs in genomic laboratory experiments have led to a steady increase of genomic feature and variation data. Multiple genomic data formats exist for sharing these data, and whilst they are similar, they are addressing slightly different data viewpoints and are consequently not fully compatible with each other. The fragmentation of data format specifications makes it hard to integrate and interpret data for further

Higgs bosons, mars missions, and unicorn delusions: How to deal with terms of dubious reference in scientific ontologies

Year: 2011 Venue: Proceedings of the Second International Conference on Biomedical Ontology Authors: Stefan Schulz, Mathias Brochhausen, Robert Hoehndorf Abstract Realist ontologies claim to represent what exists. Scientific discourse, however, often contains non-referring terms when describing hypotheses, plans, or ideas. We present a framework in which a realist ontology is embedded in an description logics theory, which is indifferent regarding the existence of class members, and which may include representational units for representing various kinds of non-referring terms. Using a taxonomy

Hyaline Arteriolosclerosis in 30 Strains of Aged Inbred Mice

Year: 2019 Venue: Veterinary Pathology Authors: Timothy K. Cooper, Kathleen A. Silva, Victoria E. Kennedy, Sarah M. Alghamdi, Robert Hoehndorf, Beth A. Sundberg, Paul N. Schofield, John P. Sundberg DOI: 10.1177/0300985819844822 Abstract During a screen for vascular phenotypes in aged laboratory mice, a unique discrete phenotype of hyaline arteriolosclerosis of the intertubular arteries and arterioles of the testes was identified in several inbred strains. Lesions were limited to the testes and did not occur as part of any renal, systemic, or pulmonary arteriopathy or vasculitis phenotype

Identifying aberrant pathways through integrated analysis of knowledge in pharmacogenomics

Year: 2012 Venue: Bioinformatics Authors: Robert Hoehndorf, Michel Dumontier, Georgios V. Gkoutos Abstract Motivation: Many complex diseases are the result of abnormal pathway functions instead of single abnormalities. Disease diagnosis and intervention strategies must target these pathways while minimizing the interference with normal physiological processes. Large scale identification of disease pathways and chemicals that may be used to perturb them requires the integration of information about drugs, genes, diseases and pathways. This information is currently distributed over several

Improved characterisation of clinical text through ontology-based vocabulary expansion

Year: 2021 Venue: Journal of Biomedical Semantics Authors: Luke T. Slater, William Bradlow, Simon Ball, Robert Hoehndorf, Georgios V Gkoutos DOI: 10.1186/s13326-021-00241-5 Abstract Ontology-based phenotype profiles have been utilised for the purpose of differential diagnosis of rare genetic diseases, and for decision support in specific disease domains. Particularly, semantic similarity facilitates diagnostic hypothesis generation through comparison with disease phenotype profiles. However, the approach has not been applied for differential diagnosis of common diseases, or generalised

Improving the classification of cardinality phenotypes using collections

Year: 2023 Venue: Journal of Biomedical Semantics Authors: Sarah M. Alghamdi, Robert Hoehndorf DOI: 10.1186/s13326-023-00290-y Abstract We reformulate the phenotypes of collections of entities using an ontological theory of collections. By reformulating phenotypes of collections in phenotypes ontologies, we avoid potentially incorrect inferences pertaining to the cardinality of these collections. We apply our method to two phenotype ontologies and show that the reformulation not only removes some problematic inferences but also quantitatively improves biological data analysis. Topics Applied

In silico exploration of Red Sea Bacillus genomes for natural product biosynthetic gene clusters

Year: 2018 Venue: BMC Genomics Authors: Ghofran Othoum, Salim Bougouffa, Rozaimi Razali, Ameerah Bokhari, Soha Alamoudi, Andr\'e Antunes, Xin Gao, Robert Hoehndorf, Stefan T. Arold, Takashi Gojobori, Heribert Hirt, Ivan Mijakovic, Vladimir B. Bajic, Feras F. Lafi, Magbubah Essack Abstract The increasing spectrum of multidrug-resistant bacteria is a major global public health concern, necessitating discovery of novel antimicrobial agents. Here, members of the genus Bacillus are investigated as a potentially attractive source of novel antibiotics due to their broad spectrum of antimicrobial

In silico screening for candidate chassis strains of free fatty acid-producing cyanobacteria

Year: 2017 Venue: BMC Genomics Authors: Olaa Motwalli, Magbubah Essack, Boris R. Jankovic, Boyang Ji, Xinyao Liu, Hifzur Rahman Ansari, Robert Hoehndorf, Xin Gao, Stefan T. Arold, Katsuhiko Mineta, John A. C. Archer, Takashi Gojobori, Ivan Mijakovic, Vladimir B. Bajic DOI: 10.1186/s12864-016-3389-4 Abstract Finding a source from which high-energy-density biofuels can be derived at an industrial scale has become an urgent challenge for renewable energy production. Some microorganisms can produce free fatty acids (FFA) as precursors towards such high-energy-density biofuels. In particular

Integrating phenotype ontologies with PhenomeNET

Year: 2016 Venue: Proceedings of Ontology Matching Workshop 2016 Authors: Miguel Rodriguez-Garcia, Georgios V. Gkoutos, Paul N. Schofield, Robert Hoehndorf Abstract PhenomeNET is a system for disease gene prioritization that includes as one of its components an ontology designed to integrate phenotype ontologies. While not applicable to matching arbitrary ontologies, PhenomeNET can be used to identify related phenotypes in different species, including human, mouse, zebrafish, nematode worm, fruit fly, and yeast. Here, we apply the PhenomeNET to identify related classes from four phenotype and

Interactively Exploring Graph Coloring Algorithms in a Bilingual Web Platform with Gamification

Year: 2017 Venue: Proceedings of EdMedia: World Conference on Educational Media and Technology 2017 Authors: Maha Alrashed, Lujain Alharbi, Omamah Talal Al-Muhammadi, Salha Bahadiq, Robert Hoehndorf, Liam Mencel Abstract Graph coloring is a concept in graph theory that has many real world applications, such as scheduling and map coloring, thus making it an essential part of a computer science curriculum. Most graph theory courses are taught using standard methods such as with textbooks or a blackboard. Such methods introduce graph theory without providing the student with an adequate

Interoperability between biomedical ontologies through relation expansion, upper-level ontologies and automatic reasoning

Venue: PLOS ONE Authors: Robert Hoehndorf, Michel Dumontier, Anika Oellrich, Dietrich Rebholz-Schuhmann, Paul N. Schofield, Georgios V. Gkoutos Abstract Researchers design ontologies as a means to accurately annotate and integrate experimental data across heterogeneous and disparate data- and knowledge bases. Formal ontologies make the semantics of terms and relations explicit such that automated reasoning can be used to verify the consistency of knowledge. However, many biomedical ontologies do not sufficiently formalize the semantics of their relations and are therefore limited with respect

Interoperability between phenotype and anatomy ontologies

Venue: Bioinformatics Authors: Robert Hoehndorf, Anika Oellrich, Dietrich Rebholz-Schuhmann Abstract Motivation: Phenotypic information is important for the analysis of the molecular mechanisms underlying disease. A formal ontological representation of phenotypic information can help to identify, interpret and infer phenotypic traits based on experimental findings. The methods that are currently used to represent data and information about phenotypes fail to make the semantics of the phenotypic trait explicit and do not interoperate with ontologies of anatomy and other domains. Therefore

Investigation of the fundamental strategy for interoperability of description of biological measurements

Year: 2011 Venue: Proceedings of the Second International Conference on Biomedical Ontology Authors: Hiroshi Masuya, Georgios V. Gkoutos, Nobuhiko Tanaka, Kazunori Waki, Yoshihiro Okuda, Tatsuya Kushida, Norio Kobayashi, Koji Doi, Kouji Kozaki, Robert Hoehndorf, Shigeharu Wakana, Tetsuro Toyoda, Riichiro Mizoguchi Abstract Aiming the facilitation of the advanced integration of measurement data across various biological experiments, we have investigated the fundamental methodology to expand the Phenotypic Quality Ontology (PATO) commonly used for descriptions of biological phenotypes with the

Klarigi: Characteristic explanations for semantic biomedical data

Year: 2023 Venue: Computers in Biology and Medicine Authors: Luke T. Slater, John A. Williams, Paul N. Schofield, Sophie Russell, Samantha C. Pendleton, Andreas Karwath, Hilary Fanning, Simon Ball, Robert Hoehndorf, Georgios V. Gkoutos DOI: 10.1016/j.compbiomed.2022.106425 Abstract Annotation of biomedical entities with ontology classes provides for formal semantic analysis and mobilisation of background knowledge in determining their relationships. To date, enrichment analysis has been routinely employed to identify classes that are over-represented in annotations across sets of groups, such

Large-Scale Knowledge Integration for Enhanced Molecular Property Prediction

Year: 2024 Venue: Neural-Symbolic Learning and Reasoning Authors: Yasir Ghunaim, Robert Hoehndorf DOI: 10.1007/978-3-031-71170-1_10 Abstract Pre-training machine learning models on molecular properties has proven effective for generating robust and generalizable representations, which is critical for advancements in drug discovery and materials science. While recent work has primarily focused on data-driven approaches, the KANO model introduces a novel paradigm by incorporating knowledge-enhanced pre-training. In this work, we expand upon KANO by integrating the large-scale ChEBI knowledge

Large-Scale Reasoning over Functions in Biomedical Ontologies

Year: 2016 Venue: Formal Ontology in Information Systems Authors: Robert Hoehndorf, Liam Mencel, Georgios V. Gkoutos, Paul N. Schofield Abstract A large number of biomedical resources have been developed to represent the functions of biological entities, and these resources are widely used for data integration and analysis. Expressing functions in biomedical ontologies currently uses formal representation patterns that renders basic reasoning tasks to fall in complexity classes beyond polynomial time, thereby limiting the potential of using knowledge-based methods for data integration

Lattice-Based ALC Ontology Embeddings With Saturation

Year: 2025 Venue: Neurosymbolic Artificial Intelligence Authors: Fernando Zhapa-Camacho, Robert Hoehndorf DOI: 10.1177/29498732251340186 Abstract Generating vector representations (embeddings) of OWL ontologies is a growing task due to its applications in predicting missing facts and knowledge-enhanced learning in fields such as bioinformatics. The underlying semantics of OWL ontologies are expressed using Description Logics (DLs). Initial approaches to generate embeddings relied on constructing a graph out of ontologies, neglecting the semantics of the logic therein. Recent semantic

Lattice-Preserving ALC Ontology Embeddings

Year: 2024 Venue: Neural-Symbolic Learning and Reasoning Authors: Fernando Zhapa-Camacho, Robert Hoehndorf DOI: 10.1007/978-3-031-71167-1_19 Abstract Generating vector representations (embeddings) of OWL ontologies is a growing task due to its applications in predicting missing facts and knowledge-enhanced learning in fields such as bioinformatics. The underlying semantics of OWL ontologies is expressed using Description Logics (DLs). Initial approaches to generate embeddings relied on constructing a graph out of ontologies, neglecting the semantics of the logic therein. Recent semantic

LEP-AD: language embedding of proteins and attention to drugs predicts drug-target interactions

Year: 2026 Venue: Journal of Cheminformatics Authors: Reem Alsulami, Robert Lehmann, Anuj Daga, Sumeer A. Khan, Raik Gr\"unberg, Ahmed Abogosh, David Gomez Cabrero, Stefan T. Arold, Robert Hoehndorf, Jesper Tegner, Narsis A. Kiani DOI: 10.1186/s13321-026-01167-9 Abstract INTRODUCTION: Predicting drug-target interactions remains a significant challenge in drug development and lead optimization. Recent advances have leveraged machine learning algorithms to model drug-target interactions from molecular and sequence data. MATERIALS AND METHODS: In this work, we use Evolutionary Scale Modeling (ESM

Linking PharmGKB to phenotype studies and animal models of disease for drug repurposing

Venue: Pacific Symposium on Biocomputing (PSB) Authors: Robert Hoehndorf, Anika Oellrich, Dietrich Rebholz-Schuhmann, Paul N. Schofield, Georgios V. Gkoutos Abstract The investigation of phenotypes in model organisms has the potential to reveal the molecular mechanisms underlying disease. The large-scale comparative analysis of phenotypes across species can reveal novel associations between genotypes and diseases. We use the PhenomeNET network of phenotypic similarity to suggest genotype--disease association, combine them with drug--gene associations available from the PharmGKB database, and

LLM Agent Based Protein Function Prediction

Year: 2025 Venue: Biocomputing 2026 Authors: Fernando Zhapa-Camacho, Olga Mashkova, Robert Hoehndorf, Maxat Kulmanov DOI: 10.1142/9789819824755_0036 Abstract Generating vector representations (embeddings) of OWL ontologies is a growing task due to its applications in predicting missing facts and knowledge-enhanced learning in fields such as bioinformatics. The underlying semantics of OWL ontologies are expressed using Description Logics (DLs). Initial approaches to generate embeddings relied on constructing a graph out of ontologies, neglecting the semantics of the logic therein. Recent

Logical Gene Ontology Annotations (GOAL): exploring gene ontology annotations with OWL

Venue: Journal of Biomedical Semantics Authors: Simon Jupp, Robert Stevens, Robert Hoehndorf Abstract MOTIVATION:Ontologies such as the Gene Ontology (GO) and their use in annotations make cross species comparisons of genes possible, along with a wide range of other analytical activities. The bio-ontologies community, in particular the Open Biomedical Ontologies (OBO) community, have provided many other ontologies and an increasingly large volume of annotations of gene products that can be exploited in query and analysis. As many annotations with different ontologies centre upon gene products

Media coverage

Selected press coverage and KAUST/CEMSE highlights about the Bio-Ontology Research Group. 2022 · KAUST CEMSE Appointment of Prof. Xin Gao as Interim Director and Prof. Robert Hoehndorf as Interim Associate Director of CBRC 2020 · KAUST CEMSE Robert Hoehndorf promoted to associate professor at KAUST 2018 · KAUST CEMSE KAUST and KACST join forces to prevent infectious diseases 2018 · KAUST CEMSE Assistant Professor Robert Hoehndorf awarded Center Partnership Fund 2016 · KAUST CEMSE Professor Robert Hoehndorf's PhenomeNET method wins at Ontology matching workshop For weekly group news see the

Microbial communities

Microbial communities drive most of the biogeochemical and biomedical processes that sustain life, yet their functional repertoire remains poorly characterised: metagenomic samples are dominated by proteins with no close homolog in curated databases, and most prediction tools have been trained on eukaryotic sequences. Our work in this area lifts single-protein function prediction up to the level of whole microbial communities by combining ontology-aware deep learning with multi-scale systems analysis. The distinctive angle is to treat metagenomes not as bags of genes but as functional systems

Molecular basis and cellular effects of Janus-class–driven cytoplasmic PYK2 coacervates

Year: 2026 Venue: Communications Biology Authors: Giovanni Colombo, Israa Salem, Kacper Szczepski, Piao Yu, Shaden Alfaiyz, Francisco Javier Guzman-Vega, Ahmed Abogosh, Maxat Kulmanov, Samah Al-Harthi, Gress Kadare, Robert Hoehndorf, Jean-Antoine Girault, Łukasz Jaremko, Afaque A. Momin, Stefan T. Arold DOI: 10.1038/s42003-025-09463-0 Abstract Kinase activity is increasingly linked to biomolecular phase separation. Focal adhesion kinase (FAK) forms membrane-associated condensates with paxillin to promote adhesion. Here we show that its paralogue, proline-rich tyrosine kinase 2 (PYK2)

Mouse genetic and phenotypic resources for human genetics

Venue: Human Mutation Authors: Paul N. Schofield, Robert Hoehndorf, Georgios V. Gkoutos Abstract The use of model organisms to provide information on gene function has proved to be a powerful approach to our understanding of both human disease and fundamental mammalian biology. Large-scale community projects using mice, based on forward and reverse genetics, and now the pan-genomic phenotyping efforts of the International Mouse Phenotyping Consortium (IMPC), are generating resources on an unprecedented scale which will be extremely valuable to human genetics and medicine. We discuss the nature

Mouse model phenotypes provide information about human drug targets

Year: 2013 Venue: Bioinformatics Authors: Robert Hoehndorf, Tanya Hiebert, Nigel W. Hardy, Paul N. Schofield, Georgios V. Gkoutos, Michel Dumontier Abstract Motivation: Methods for computational drug target identification utilize information from diverse information sources to predict or prioritize drug targets for known drugs. One set of resources that has been relatively neglected for drug repurposing are animal model phenotypes.Results: We investigate the use of mouse model phenotypes for drug target identification. To achieve this goal, we first integrate mouse model phenotypes and drug

mOWL

Python library for machine learning with biomedical ontologies. Unifies projection-, axiom- and geometric-embedding methods (EL Embeddings, ELBE, BoxSquaredEL, OWL2Vec*, DL2Vec, OPA2Vec) behind one API, with first-class OWLAPI access and PyTorch integration. Get it GitHub: https://github.com/bio-ontology-research-group/mowl ★ 91 Homepage: https://mowl.readthedocs.io Developed in projects IBNSINA-QI: Integrating Biomedical Networks and Semantic Information for Neural network Analysis of Quantitative Information Towards sound, complete, and explainable machine learning with biomedical ontologies

Multi-faceted semantic clustering with text-derived phenotypes

Year: 2021 Venue: Computers in Biology and Medicine Authors: Luke T. Slater, John A. Williams, Andreas Karwath, Hilary Fanning, Simon Ball, Paul N. Schofield, Robert Hoehndorf, Georgios V. Gkoutos DOI: 10.1016/j.compbiomed.2021.104904 Abstract Identification of ontology concepts in clinical narrative text enables the creation of phenotype profiles that can be associated with clinical entities, such as patients or drugs. Constructing patient phenotype profiles using formal ontologies enables their analysis via semantic similarity, in turn enabling the use of background knowledge in clustering

Nail abnormalities identified in an ageing study of 30 inbred mouse strains

Year: 2019 Venue: Experimental Dermatology Authors: Sarah C. Linn, Allison M. Mustonen, Kathleen A. Silva, Victoria E. Kennedy, Beth A. Sundberg, Lesley S. Bechtold, Sarah M. Alghamdi, Robert Hoehndorf, Paul N. Schofield, John P. Sundberg DOI: 10.1111/exd.13759 Abstract In a large-scale ageing study, 30 inbred mouse strains were systematically screened for histologic evidence of lesions in all organ systems. Ten strains were diagnosed with similar nail abnormalities. The highest frequency was noted in NON/ShiLtJ mice. Lesions identified fell into two main categories: acute to chronic

Nanodesigner: resolving the complex-CDR interdependency with iterative refinement

Year: 2025 Venue: Journal of Cheminformatics Authors: Melissa Maria Rios Zertuche, Senay Kafkas, Dominik Renn, Magnus Rueping, Robert Hoehndorf DOI: 10.1186/s13321-025-01069-2 Abstract Abstract Camelid heavy-chain only antibodies consist of two heavy chains and single variable domains (VHHs), which retain antigen-binding functionality even when isolated. The term “nanobody” is now more generally used for describing small, single-domain antibodies. Several antibody generative models have been developed for the sequence and structure co-design of the complementarity-determining regions (CDRs)

Neural Multi-hop Logical Query Answering with Concept-Level Answers

Venue: The Semantic Web – ISWC 2023 Authors: Zhenwei Tang, Shichao Pei, Xi Peng, Fuzhen Zhuang, Xiangliang Zhang, Robert Hoehndorf Abstract Neural multi-hop logical query answering ({LQA}) is a fundamental task to explore relational data such as knowledge graphs, which aims at answering multi-hop queries with logical operations based on distributed representations of queries and answers. Although previous {LQA} methods can give specific instance-level answers, they are not able to provide descriptive concept-level answers, where each concept is a description of a set of instances. Concept