Multi-Omics & Biostatistics for Predictive Biomarkers in Chronic Diseases

We offer potential partners complementary expertise in biomarker discovery, transcriptomics and multi-omics, AI/XAI-based predictive modelling, biostatistics and clinical research methodology for chronic non-communicable diseases, with a particular focus on diabetes and metabolic syndrome. Importantly, through collaboration with a Department of Internal Medicine, we have access to clinical data and patient-derived biological samples, mainly from patients with diabetes and metabolic syndrome. This provides a direct translational setting for linking molecular profiles with clinically relevant phenotypes, treatment exposure and disease outcomes.

Clinical and translational access: Our clinical collaboration enables collection and analysis of patient-derived samples together with routine clinical and phenotypic data. The initial disease focus is diabetes and metabolic syndrome, providing a clinically relevant population for biomarker discovery and validation. These local data can be integrated with external longitudinal cohorts, biobanks and RWD to strengthen model development and independent validation.

Transcriptomics and multi-omics: Our expertise includes bulk RNA-seq, sn/scRNA-seq, differential expression, pathway analysis, dimensionality reduction, clustering, co-expression networks and multi-omics integration. We have extensive experience translating complex molecular profiles into clinically interpretable signatures. Examples include transcriptomic biomarkers stratifying biochemical recurrence beyond clinical features (PMID: 37176106) and WGCNA-based identification of a predictive WNT pathway signature (PMID: 31134156). These workflows can be adapted to blood-derived or other clinically accessible biospecimens in metabolic disease.

Systems biology and WWOX: We use GSEA, WGCNA, pathway activity analysis and network-based modelling to identify regulatory programmes or multimodal biomarkers. WWOX is of particular interest as a mechanistic candidate within broader signalling and metabolic networks, rather than as a stand-alone marker. Our previous work demonstrated _WWOX_-associated regulation in neuronal differentiation (PMID: 31543760) and _WWOX_-related TGFα-EGFR signalling (PMID: 35290621). In metabolic disease, we propose evaluating WWOX expression and its downstream network context within unbiased transcriptomic and multi-omic signatures associated with insulin resistance, metabolic dysregulation, complications and treatment response.

Biomarker discovery & XAI-supported predictive modelling: We develop prognostic and predictive signatures using multivariable modelling, machine learning, neural-network approaches, feature selection, subgroup identification and model validation. XAI is a key component, allowing identification of molecular and clinical variables driving individual predictions and improving biological interpretation and clinical transparency. Our methodological background also includes survival-based optimisation of continuous biomarkers through the Evaluate Cutpoints approach (PMID: 31319941). For diabetes and metabolic syndrome, these methods can support prediction of disease progression, complications and treatment response from combined clinical, molecular and RWD features.

Biostatistics and validation: We provide expertise in endpoint definition, sample-size and power calculations, Statistical Analysis Plans, missing-data strategies, multivariable and longitudinal analyses, internal and external validation, discrimination, calibration and subgroup analyses. Our experience in investigator-initiated and multicentre clinical studies provides a robust methodological framework for biomarker development and validation.

Clinical implementation, Health Technology Assessment (HTA) and Health Economics: Our scientific profile is complemented by Healthcare Management, including health economics and HTA. This supports evaluation of clinical utility, implementation pathways, scalability and potential cost-effectiveness of biomarker-based strategies, which is particularly relevant for chronic diseases requiring long-term management.

Partnership sought: We may particularly complement partners providing large, deeply phenotyped longitudinal cohorts, biobanks and RWD infrastructures with repeated measurements, treatment information and long-term follow-up. Such resources would complement our direct access to diabetes/metabolic-syndrome patients and samples and enable external validation across populations and healthcare settings.

Together, we offer an end-to-end translational workflow linking patient-derived samples and clinical data with multi-omics, systems biology, AI/XAI-supported biomarker development, rigorous biostatistical validation and assessment of clinical utility in metabolic disease.