Atul Butte on translational bioinformatics

Embed gadget

In his thoughtful and provocative 2012 TedMed talk, Atul Butte, Chief of Systems Medicine at Stanford University, discusses the field of translational bioinformatics and introduces the idea that big data is fundamentally changing the way science is being done.

The National Institutes of Health (NIH) and other organizations host mega-scale databases that include a variety of data such as gene expression measurements, sequence data, indexed text data from publications, network relationships, environmental scans, and much more. Furthermore, NIH is increasingly specifying data sharing requirements as part of funding of research.

Dr. Butte promotes a new approach to medical discovery that begins with fusing information from a variety of sources and then using machine learning to search for patterns and relationships. In data-driven science, researchers extract testable hypotheses from these relationships and then design experiments to validate them.

The commercialization of validation experiments such as gene knockouts and others, substantially reduces the cost of performing these experiments and increasingly researchers are outsourcing these experiments because they can be done much more cheaply and reliably by companies specializing in particular types of experiments with specific animal models. Outsourcing then changes the fundamental structure and economy of scientific research and more emphasis in the future will be on the discovery of patterns, the formulation of hypotheses, and the design of experiments, while many routine laboratory tasks are likely to be outsourced.

Butte also discusses the complex inter-relationship between genetics, toxicology, and the environment as being studied in the new field of environomics. Dr. Butte delivered a more extensive discussion of these ideas at the NIH in a lecture entitled: Translational Bioinformatics: Transforming 300 Billion Points of Data.