 |  | Recordings & Slide Decks | MEET OUR SPONSORS |
| Download Event Circular | Return to Event Page |
The following Keynote Speakers will preset at this Symposium on June 27th, 2018.
| Speaker | Presentation Topic | |
|---|---|---|
 | Zak Kohane, MD, PhD Professor and Chair, Department of Biomedical Informatics, Harvard Medical School | Clearing the Cloudy Forecast for Clinic-Genomic Computing Even as unprecedented volumes of clinical data are generated from electronic health records, non-institutional monitoring several substantive concerns are being raised that may impede progress in using these data to improve clinical care and accelerate research. These include: data security, data rights of patients, uncertainty about who pays for “cloud” hosting, proliferation of heterogeneous clinical data repositories, and shifting expectations of which data and analyses have to be preserved for publication and conformance with emerging standards for reproducibility. In reviewing these challenges, I will articulate and illustrate with notable national and international examples how such obstacles are being overcome by leading groups and how their examples can be emulated. I will close with a focus on a set of technical solutions that will simplify and accelerate the policy discussions. |
 | Eric Perakslis, PhD Chief Science Officer at Datavant | Pushing Data Uphill: The Sociology of Open Tech and Open Data Over the past decade, great progress has been made in collaborative data science. Open source platforms such as i2b2 and tranSMART have flourished and provided low cost entry into previously complex biomedical data use cases. Despite these and other successes, medical data has not become the vibrant circulatory system of healthcare and biomedical product development that was promised. Instead, the system remains slow and complex and it is not always clear if the massive investments on federally mandated digital medical records have helped or hindered. Challenges of cost, access, completeness, interoperability, security and privacy are on the rise and brave new approaches that can push through current technological and compliance orthodoxies are needed now. |
 | George Church, PhD Professor, Harvard Medical School | Open Source integration of Omics and other precision medicine data from 2005 to 2025 PersonalGenomes.org and OpenHumans.org are demonstrations of the feasibility and utility of fully open sharing (Wikipedia level) of highly integrated (hence identifiable) human biology data. Examples include: 1. NIST+FDA Genomeinabottle.org distributable DNA standards, 2. Coriell sharable cell lines (iPSC, B-cells, fibroblasts), 3. ENCODE and Human Cell Atlas (HCA) projects defining epigenetics of isogenic diverse cell types. 4. Critical Assessment of Genome Interpretation (CAGI, genomeinterpretation.org). 5. Re-identification studies (Sweeney and Erlich) |
 | John Halamka, MD, MS HMS Professor of International Healthcare Innovation | Emerging Models of Data Flow - APIs, Blockchain and Cloud In 2017 and 2018, three major technologies are empowering novel analytics - FHIR/APIs for fluid data exchange with EHRs, the rise of cloud hosted machine learning services, and an explosion of mobile applications. Regulatory changes such as 21st Century Cures, TEFCA, and Trump-era interoperability efforts are creating an urgency to change. In this presentation we’ll review the key technology and policy changes that are accelerating novel forms of interoperability and agile functionality. |
 | Becky Steck U Michigan Health System | Kanban and Kidneys: using Agile software development methodologies to deliver solutions to renal translational research communities Academic research institutions are no strangers to software creation, but many academic software solutions are hampered by an absence of process and code standards, scarcity of documentation, and lack of consistent resources for maintenance, support, and improvement. These issues ultimately make it difficult for these software products to fulfill their promise. I will discuss how the software development team at the University of Michigan’s Department of Nephrology applies Agile software development methodologies to both create new tools and leverage the open source tranSMART platform to provide the renal community with stable, user-centered solutions to some of their translational research needs. |
 | Shawn Murphy, MD, Ph D Partners Healthcare | The integration of i2b2 with the future of Big Data The i2b2 platform has evolved to handle many types of Big Data. The integration of Genomics, Imaging (mi2b2), and digital data is performed with simple facts and pointers. Assuring that essential classifiers are entangled with the Big Data makes it searchable and computable in multiple research integration platforms. I2b2 and Jupyter “a perfect marriage” Database languages and programming languages live side-by-side in creating phenotypes, making disease predictions, and supporting decisions regarding patient care. The use of a Jupyter notebook with i2b2 in a dynamic computing environment makes a more fluid and reproducible analysis environment. The privacy, software language, and memory support issues were worked out by wrapping the environment in a data enclave. |
 Paul Avillach | Paul Avillach, MD, PhD Department of Biomedical Informatics, Harvard Medical School | The first public version of the i2b2/tranSMART platform The real value in biomedical research lies not in the scale of any single source of data, but in the ability to integrate and interrogate multiple complementary datasets simultaneously. i2b2/tranSMART 18-1 platform enables translational bioinformatics by integrating multiple heterogeneous sources of clinical, genomics and environmental data in a meaningful way. The scalable BD2K PIC-SURE metaAPI enables combining data across different scales and resolutions, and allows integration to other platforms including JupyterHUB to drive reproducible science. i2b2/tranSMART 18-1 Beta release is here: https://github.com/i2b2-tranSMART ONE script docker-compose file creates the full interconnected easy deployment of docker images of i2b2/tranSMART 18-1 Beta release including: a) the latest version of i2b2 in production (1.7.09c), b) the latest version of tranSMART in production (16.2), c) the BD2K PIC-SURE metaAPI http://bd2k-picsure.hms.harvard.edu d) CDC NHANES public dataset including 41,474 patients already loaded in i2b2/tranSMART https://nhanes.hms.harvard.edu e) Fractalis plugin (next version of SmartR plugin) f) Auth0.com, a global leader in Identity-as-a-Service (IDaaS) integration as an option. (Script available end of April 2018) |
  | Mary Serpan and Mike Mendis | INVcheck INVcheck is an application using clinical data from ERHs, that enable emergency healthcare providers and emergency responders to pre- and post- event identify ‘at-risk’ patients who may be medically vulnerable to the impact of a natural disaster. |
 | Griffin Weber, Ph D Department of Biomedical Informatics, Harvard Medical School | Technology Demos Millisecond queries, ontology visualizations, and web client architectures |