The Computational and Systems Medicine research groups led by Professor Jeremy Nicholson and Elaine Holmes lead the world in metabolic phenotyping approaches to understanding patient stratification and public healthcare problems linking metabolism to disease risk. In clinical medicine optimisation of the patient care pathway via understanding deep patient biology is a key to personalisation. The Department of Surgery and Cancer houses both the MRC-NIHR National Phenome Centre and the NIHR/BRC Clinical Phenome Centre which are large scale spectroscopy powered metabolic research centres that are unique in the world (INSERT WEB LINKS). The patient journey phenotyping paradigm (Figure 1) is scalable and translatable to all areas of medicine and multiple patient pathways are currently under examination. From this phenotyping approach a new Phenotypically Augmented Clinical Trial (PACT) design can also be derived, using systems medicine metrics to enable dynamic patient stratification that can be visualised in interactive data-rich environments. The extra layers of phenomic data in the PACT providing a predictive physiological background model to map patient responses to novel therapies.

In principle this approach could also be reverse engineered to consider “experimental animal journeys” to model human drug trials or hospital procedures. We are also developing data rich real-time diagnostic technologies (closely aligned with the Hammersmith based MRC-NIHR National Phenome Centre and the St Mary’s based Clinical Phenome Centre) for augmented clinical and surgical decision making including the i-knife, I-endoscope and related bedside spin-off technologies. Underpinning these areas would be development of Advanced Visualisation Science, including Augmented Reality and Interactive 3D data embodiment, display and modelling approaches - already embraced in the Systems Oncology and Computational Medicine initiatives in the Department of Surgery and Cancer.

Visualisation Science, Surgical Metabonomics and Engagement Centre

21st century personalised healthcare demands completely new ways of interacting with almost unimaginably complex data. Scientists, clinicians, patients and communities all face this challenge, though it affects them differently. The Imperial Institute has already identified data visualisation as core and the data output of the MRC-NIHR National Phenome facility alone has data outputs in the Petabyte/year range, which requires advanced mathematical analysis and multivariate modelling in relation to clinical and genomic data. The development, on the same site, of Engagement Science will develop a new lexicon of engagement centred around visualisation and simulation. This additionally places engagement at the 

centre of the research process, creating a bridge between researchers, the public and data.  Incorporating Engagement in this way captures the vision and inspiration behind personalised healthcare and stratified medicine.

Embodiment of Mega-variate Data: A key translational challenge in Systems Medicine is the visualization of complex omics data sets (inked to clinical, imaging and pathological metadata) to augment clinical decision-making. This requires efficient linkage of disparate mega-data sets in relational databases, multi-omic data fusion and extraction of biological and medical knowledge. We will use a range of advanced statistical spectroscopy methods- many of which were developed in our group for systems medicine purposes including our own novel pathway connectivity algorithms such as Metabonetworks. The Metabonetworks suite utilizes biomarker input data (extracted from chemometric analysis of metabolic spectroscopy sets) to calculate optimised network structures required to express all network connections of the biomarkers based on a numerical model of all possible scalar connections between all metabolites in genome selected KEGG (or other) pathways. This also uniquely allows supraorganismal symbiotic connectivities to be made between microbiome and mammalian genomes and their metabolic products. However, such modeling methods are unfamiliar to clinicians so alternative methods of model display are required to create dynamic and visually-rich relational environments into which advanced omics models can be projected and interactively examined with simultaneously projected clinical metadata. This is a highly novel and innovative part of the proposal viz. the use of 3-D augmented reality instruments to provide data embodiment- these include existing commercial devices such as i-Dome and i-Cave that are capable of delivering multimodal data fusion models of complex worlds. These systems have multiple layers of model output possibilities expressed in the form of embedded stratified medicine navigation engines (STRATNAVS) which include decision making environment models, clinical training models and even levels for patient experience and engagement (MY-STRATNAV). This will be achieved using the same data input files and a series of kernel algorithms with different levels of metadata linkage allowing multi-level model access.

The new computational platforms will centre on layered Patient Journey or Stratified Medicine Navigational Models (STRATNAV) that can be used for multiple purposes (Figure 2), specifically:

1. To augment clinical decision-making via visualisation of the total patient journey metrics and models, to create practical knowledge for Doctors and Surgeons to use.
2. To provide a new systems-medicine training environments for clinicians and clinician scientists to enable future decision-making pathways.
3. To provide new undergraduate student training capabilities based on visualisation of systems medicine exemplar navigated patient pathways.
4. To enable patient-centred interrogation of the their own hospital journeys thus enhancing the patient experience and understanding of their particular condition and treatments- this gives the individual view of personalised healthcare. (MY-STRATNAV) This is effectively personalising the patient experience journey- and could easily be expressed in the form of an iPad application for example.
5. To provide new mechanisms of public engagement with medical science via user-friendly models of personalisation of healthcare. This would enable enhanced communication of the science of personalised healthcare models with future funders, politicians, the media and the public.