Translational Medicine in the Age of Data

Billions of clinical measurements are recorded every day and stored in electronic health systems around the world. Each one of these experiments is a window into the human system, creating the most comprehensive and diverse medical data set ever imagined. Unfortunately, traditional statistical techniques were not developed to handle such diversity, instead they excel at analyzing homogenous data sets with first order effects. Because of this, these techniques are simply unable to untangle the sophisticated web of biological pathways and genetic interactions governing the human system.

With enormous data come enormous opportunity

Data Science is a new field dedicated to developing the methods, algorithms, and tools to unravel the complexities of enormous data. In our lab we advance data science by designing rigorous computational and mathematical methods that address the fundamental challenges of health data science. Foremost, we integrate our medical observations with systems and chemical biology models to not only explain drug effects, but also further our understanding of basic biology and human disease.

One particular area of interest is the integration of high-throughput data capture technologies, such as next-generation genome and transcriptome sequencing, metabolomics, and proteomics, with the electronic medical record to study the complex interplay between genetics, environment, and disease.

For a more in-depth information on our research areas of interest see our reviews in WIREs System Biology and Medicine, Science Translational Medicine, and Clinical Pharmacology & Therapeutics.

News and Events

New case studies corroborate drug-drug interaction discovery

In a letter to the editor at JACC, clinician-researchers report on two clinical observations providing further evidence that ceftriaxone and lansoprazole increase arrhythmia risk, "via specific (not class-related) QT-prolonging effects exerted by these 2 molecules when combined. The potential harmfulness of this association should be carefully kept in mind." Read the whole letter and our response.

Scientists see progress in identifying deadly drug interactions
By Sam Roe

"The results and their interpretation," wrote Dr. Dan Roden, a leading expert on cardiac arrhythmia at Vanderbilt University who was not involved in the research, "provide important lessons for investigators interested in using 'big data' approaches to study (adverse drug reactions), other drug effects and indeed, many other aspects of the human condition." Read the whole story.

Animal Venom Database Could Be Boon To Drug Development
By Emily Mullin

"The bite of a poisonous snake, scorpion or other venomous creature could very well kill you, but it also might be able to heal certain medical conditions like cancer, diabetes and heart failure. That's the idea behind VenomKB, short for Venom Knowledge Base, the first online database that aims to catalog all the known animal toxins and their physiological effects on humans." Read the whole story.

Featured publications

Tal Lorberbaum, Kevin J. Sampson, Jeremy B. Chang, Vivek Iyer, Raymond L. Woosley, Robert S. Kass, Nicholas P. Tatonetti
Coupling Data Mining and Laboratory Experiments to Discover Drug Interactions Causing QT Prolongation.
Journal of the American College of Cardiology. Oct 2016. Source.

Yun Hao and Nicholas P. Tatonetti
Predicting G protein-coupled receptor downstream signaling by tissue expression.
Bioinformatics. July 2016. Source.

See more publications.


Our lab is in the Department of Biomedical Informatics at Columbia University as well as the Department of Systems Biology, and the Department of Medicine. We are a member of the Data Science Institute at Columbia.

Potential graduate students should apply to the Department of Biomedical Informatics Training Program or the Computational Biology Training Program at Columbia.