Trends in Genomic Data and Big Data Analytics

Scott Kahn, PhD
Vice President, Commercial
Enterprise Informatics
Illumina, Inc.


The ever-increasing output of genome sequencing instruments has recently resulted in a sub $1,000 genome inclusive of the informatics to map the reads and call variants. As researchers switch their focus from single genome analysis to cohort-level genomic analysis, and with more comprehensive genomic characterization of each subject (eg, DNA, transcriptome, epigenome, microbiome, etc), the informatics challenges are moving into the realm of Big Data analytics. This presentation will put these trends into context and will provide a perspective on how these challenges are being addressed and where future improvements are needed.



Steven Steinhubl, Scripps Healthcare (Director of Digital Health)


Wearable Sensors: Moving from the Quantified Self to the Understood Self

Through progressively miniaturized and increasingly powerful mobile computing capabilities, individuals now have the capability to monitor, track and transmit important health metrics continuously and in real time. Taking advantage of this, a wide spectrum of novel technologies has been developed to allow for personalized wellness, acute disease diagnostics, and chronic condition management from home that would otherwise have required an office or hospital visit. Wrist-worn sensor technologies currently available or in late-stage development may best exemplify both the capabilities and analytic challenges these remarkable advances are bringing to the healthcare setting. Over a dozen important health and wellness parameters are capable of being monitored continuously with a watch-like device during routinely daily activities; parameters currently only available in an Intensive Care Unit setting such as beat-to-beat blood pressure, cardiac output, ECG, oxygen saturation, and more. Beyond being just a more convenient way for vital signs to be monitored, these multiple, continuous data streams offer tremendous opportunities to understand an individual's unique and personalized physiologic responses to daily stressors, and most importantly, help guide healthy responses to them. Transforming these numerous, vast and inter-related data streams into understandable and actionable information for the individual and their healthcare team is a critical requirement for mobile sensor technology to achieve its potential to improve the health and wellness of all of us.



Balaji Krishnapuram, Siemens Healthcare (Head of Research, Health Services Innovation Center)

Rapid Learning using Privacy-Preserving Distributed Data-Mining


Recent advances in technology and data acquisitions costs are leading to an explosion of electronic data for bioinformatics and clinical research. At the same time we are witnessing a new paradigm for clinical research based on secondary use of data from Electronic Medical Records (EMR). In this talk, we describe a Health IT system that supports Rapid Learning across hospitals in Germany, Belgium and Netherlands. We describe that technological developments that enable us to conduct clinical research by bringing the computation to the data in federated databases in each hospital. The system avoids centralizing the data and preserves patient privacy, thus overcoming ethical, political and legal challenges to sharing patient data. Further, it uses ontologies and machine learning methods to dramatically reduce administrative and IT costs for collecting, normalizing and exchanging information across disparate source systems that use different languages, clinical protocols, database schema. We demonstrate the impact of the system on translational clinical research based on a case study across 5 hospitals in 4 countries.