Personalized precision medicine (PPM) is rapidly emerging as the future in several areas of clinical diagnostics, including cancer. The advent of PPM is dependent on recently developed capability to (a) perform comprehensive and costeffective molecular phenotyping, often through sequencing, and (b) ability transform large datasets into knowledge and actionable information using advanced statistical and machine learning methodologies. Development of predictive and prognostic PPM models that combine multiple types of molecular and clinical data for patient stratification in respect to prognosis, molecular subtype and probability of treatment response is a nontrivial challenge both in respect to developing and evaluating specific models and in developing the underlying statistical and computational approaches.
In my research group we apply modern statistical and machine learning methods to turn large molecular dataset into insights and predictive patient stratification models for PPM applications in cancer, currently with applications in breast cancer (BC), acute myeloid leukemia (AML) and prostate cancer (PC). Driven by our research in precision medicine we develop novel statistical and bioinformatic methodologies and strategies to address challenges in this research domain. In the area of singlecell molecular profiling of intratumor heterogeneity we also lead experimental studies generating primary datasets. Due to the crossdisciplinary nature of PPM projects we regularly collaborate closely with other domain experts in molecular profiling and sequencing, oncology, pathology and epidemiology. Our work contributes to the development of PPM and has the potential to substantially improve patient outcomes by providing patients treatment options with high probability of response and by reducing over and undertreatment.
Principal areas of investigation: Theme A. Development and validation of novel molecularbased PPM models for improved patient stratification (prognostic and predictive) using DNA and RNAsequencing together with statistical and machine learning methods. Theme B. Development of nextgeneration singlecell PPM methodologies and models accounting for for intratumour heterogeneity at the singlecell level. Theme C. Development of artificial intelligence (deeplearning) for automated histopathology image classification in prostate cancer and integrative analyses with DNA and RNAsequencing data