Adriano Martinelli 
contact information
links
profile
The main objective of my work is to analyse tumour heterogeneity in various omics modalities and deepen our understanding of how this heterogeneity influences patient outcome and therapeutic response.
Tumor Heterogeneity
My work on tumor heterogeneity focuses on the development of methods that quantify variability in omic profiles, cell-cell interactions, entropic measures that quantify the phenotypic diversity in a tumour micro-environment or immune infiltration patterns. On a cohort level, I try to use these measures in machine learning models to discover correlations between tumour heterogeneity and patient response.
Drug Response Modelling
My modelling efforts are also directed towards drug response predictions based on different omics data. Whereas traditional deep learning models have been extensively employed for predictive tasks in drug perturbation, only very little research with respect to spatial tumour heterogeneity has been published. To this end, we will explore recently developed graph representational learning methods that enable an entity-level modelling of the data and thus circumvent limitations of other deep learning model architectures.
Accelerate Research Discovery
The developed (software) tools are made accessible to the broader research community to accelerate research and drive further insight into the intricacies of tumour heterogeneity.