FEB 24, 2016 10:00 AM PST

Harnessing the Power of Healthcare Data with Causal Machine Learning to drive Personalized and Effective Treatment Interventions

  • Co-founder and executive vice president, GNS Healthcare
      Iya Khalil, co-founder and executive vice president of GNS, has promoted use of the Cambridge, MA-based company's technology in the biopharma industry. One of the inventors of the company's computational platform, her work has helped major players such as Biogen Idec and Johnson & Johnson to make computer-driven discoveries for their respective R&D programs.

      Biopharma companies' lack of biological understanding explains a great many miserable failures in drug development. GNS' approach, known as reverse engineering/forward simulation, is able to crunch billions of genomic, molecular and clinical data points and identify cause-and-effect relationships. Such insights help pinpoint predictive biomarkers for experimental drugs and new targets for therapies.

      Khalil has pushed personalized medicine toward reality for more than a decade. She and GNS CEO Colin Hill started the company more than 11 years back while Khalil was wrapping up her Ph.D. in theoretical physics at Cornell University, and she has been at the forefront of integrating computational biology into the drug development process.


    Every day, vast amounts of healthcare data are collected from clinical trials as well as real world medical visits on patient treatment regimens and subsequent clinical outcomes. This big data raw material provides a rich asset to investigate for understanding therapeutic effectiveness and patient care. Datasets range from genomics and other ‘omic data to clinical to medical and pharmacy claims to electronic medical records to registries and beyond, and analysis can discover and predict biomarkers, drivers of disease, novel interventions, mechanisms of action, drug combinations, disease models, portfolio optimization, and personalized care/treatment algorithms. Key to leveraging this data and uncovering which treatments and interventions specifically improve a patient’s health, are powerful analytic approaches. Utilizing causal mathematics and machine learning to create in silico disease networks directly from data has been a successful approach to identify predictive and causal mechanistic associations. Simulations of resultant models unlock the knowledge within complex data, enabling personalized, actionable predictions and precision targeting of interventions. Fully realizing the power of precision medicine to identify and predict patient outcomes will significantly increase the ability of healthcare leaders and professionals to make better decisions to improve patient care.

    The seminar will focus on the following learning objectives: [1] case studies detailing the data and experimental design that have yielded success and [2] key actionable insights generated from models and analytics that can be leveraged in drug discovery and development all the way to healthcare patient setting.

    Show Resources
    Loading Comments...