A Learning Healthcare System

Screen Shot 2015-08-09 at 10.33.19 AMProfessor Russ Altman, department chair and professor of bioengineering, has always worked to harness the power of computing to understand our biological systems. Humans respond to drugs differently, and Altman’s team uses our genetic information to understand why.  By combining a biological understanding of the proteins with the physical behavior of the molecules in a simulated environment, Altman’s team can make critical predictions about individual response to drugs. Then, they can use this information to design new drugs to target diseases.

Altman has always felt strongly about conducting interdisciplinary research, undertaking numerous initiatives from studying protein structure and function (proteomics), to understanding how differences in our genes change our responses to certain drugs (pharmacogenomics). Altman is also a principal investigator of Simbios, an initiative that runs physical simulations of biological systems. Simbios allows scientists to understand the dynamics and underlying biological mechanisms for even large atom systems.

Professor Altman’s team is trying to answer some of the most fundamental questions in current pharmaceutical industries by studying pharmacogenomics. How do we characterize how good a medicine is? How do we determine which drugs to prescribe an individual? As Altman explained, the genes we inherit from our parents play a crucial role in determining how we respond to various drugs and medications.

Personally tailoring each patient’s course of treatment maximizes the efficacy of drugs, leading to higher quality healthcare. Through various computational techniques coupled with biological understanding of genetics and drugs, Altman’s team has facilitated personalized treatment by creating the first pharmacogenomics database, pharmgkb.com. The database provides valuable clinical information allowing physicians to assign a personalized prognosis to an individual’s genome.

In one case, Altman’s team was studying two widely prescribed drugs: paroxetine, an antidepressant, and pravastatin, a lipid-lowering agent. With collaborators, Altman’s team determined that when taken together, paroxetine and pravastatin lead to an increase in blood glucose that is even larger in diabetic patients. By using retrospective computational studies like these to guide clinical understanding, physicians can carefully assign medications that optimize treatment.

Prof Altman is still focused on the last translational component of his pharmacogenomics research. Given the rapid progress of computing since the 1980s, he believes that doctors need to be trained in applying the information gained from computational platforms such as pharmgkb.

According to Altman, doctors find incorporating such genetic information in their prognosis to be stressful, given time constraints.The general practitioner has an average of 12 minutes per patient.

He hopes to provide a computational platform that doctors can easily use to access clinical information. Ideally, under this “learning healthcare system” a doctor could quickly make a prognosis with an iPad or mobile device, using every patient’s medical and gene data to determine the appropriate treatment. Every patient’s data could help improve another patient’s treatment.

Altman believes that “failures in research should be looked at as a research experience.” He hopes one day to use this experience and his bioinformatics models to bring a large scale positive impact to health care industries. His work towards the steady curation of pharmacogenomics data can lead to a system where each patient has a tailored, effective treatment.


by Vishnu Shankar