Integrative analyses of molecular and clinical Data from clinical trials
My research interests are built upon computational analyses of the ever-growing rich datasets of multi-analyte measurements (genomic, genetic, proteomic, metabolites, imaging etc). More specifically, our team will lead the integrated analysis of molecular data with demographic, pathologic and outcome data to identify predictive and prognostic biomarkers for selective therapeutic agents. Using a systemic approach, my research group will develop statistical analytic tools for studying the relationship between survival data and high-dimensional genomic data, such as variable selection methods, ensemble learning and cross-data-type prediction to incorporate other data types like DNA copy number, methylation, and proteomics as available into our predictive algorithms, used for stratified medicine.
Identification of biological endpoints for emerging therapeutic targets
Response-adaptive randomised studies have been advocated as an effective way to allocate subpopulations of patients such that more patients receive better treatments. My other important research priority is to identify intermediate endpoints and develop robust assays for perioperative windows of opportunity and neoadjuvant trials. The development of adaptive trials is an important strategy for the ICR-CTSU. With Professor Judith Bliss, I will concurrently develop clinical trials for rapid evaluation of therapeutic strategies to test lead hypotheses while laying groundwork for future trials.
With Professors Bliss and Mitch Dowsett, we are developing a novel clinical research platform that will embed trials to evaluate the matching of combinations of endocrine plus targeted therapy with biomarkers of specific endocrine resistance pathways (POETIC-2). In collaboration with Professor Andrew Tutt, Dr Sheeba Irshad, and Professor Bliss, we have co-developed the “PHOENIX” trial, a UK platform initiative constituting a post neoadjuvant pre-surgical disease profiling and novel therapy “window of opportunity” biomarker endpoint trial.
Implementation of cognitive computing in clinical trials network
We will be under a deluge of data generated from the research initiatives, including sequencing and gene expression data, protein abundance by mass spectrometry, immunohistochemical images, and clinical outcome.
Leveraging strategic alliance with the clinical trials network, The National Cancer Research Institute Cellular-molecular pathology initiative, and existing collaboration with NanoString® Technologies, my research interest is to collect the genomic and protein expression profiles of tumours (e.g. applying the NanoString® 3-D biology technology) from “exceptional drug responders” to learn the underlying biology mechanisms of these outliers from clinical trials. Our group will assess various applied machine-learning methods for genomics data linked with clinical information. If successful, we may potentially be able to suggest alternative drug options for these “extreme responders”.