CHICAGO (GenomeWeb) – Vanderbilt University's new initiatives in big data and data science may be university-wide, but bioinformatics and genomic medicine in particular stand to have outsize influence. Accordingly, genomics programs at Vanderbilt University Medical School may reap major benefits from the efforts.
That is because one of the originators of the program — and co-chair of the month-old Vanderbilt Data Science Visions Working Group — is Yu Shyr, the Harold L. Moses Professor of Cancer Research, chairman of biostatistics, and director of the Vanderbilt University Medical Center's Center for Quantitative Sciences.
Plus, according to Shyr, the initial idea for a whole-university focus on data science came from Provost Susan Rae Wente, who has a background in cell biology.
"We got very positive feedback from our board members to support Vanderbilt to have this big initiative about data science," Shyr recalled. In fact, the university board gave unanimous approval to the formation of the working group. "[Senior leadership] feels that data science will be a big part of the future of education and research," Shyr said.
The multidisciplinary Data Science Visions Working Group, made up of 20 faculty members from all over the university, will identify research and educational areas that could make better use of data analytics, and highlight opportunities for collaboration among schools and departments. The group will deliver an action plan to Wente and deans of every Vanderbilt school and college in August 2018.
"This opportunity will really give us, across different schools, across different fields, a chance to work together," said Shyr, who holds professorships in biomedical informatics, cancer biology, and health policy. "Our job here is to work with all the faculty from all schools."
Even the law and music schools are represented in the Data Science Visions Working Group.
Vanderbilt has a long history in biomedical informatics, particularly in clinical informatics; VUMC built its own electronic health record starting in the 1980s. "We also have a very strong genomics group," Shyr said.
Shyr, as director of the Center for Quantitative Sciences, already works with faculty from 25 different divisions within Vanderbilt University Medical Center. These focus on complicated data science, including in genomics, proteomics, lipidomics, and imaging.
Shyr also directs Vanderbilt Technologies for Advanced Genomics Analysis and Research Design, or VANGARD, the technology core that is paired with the Vanderbilit Technologies for Advanced Genomics (VANTAGE) wet lab that handles all the genomic sequencing analysis at VUMC. VANGARD supports VANTAGE with bioinformatics and other data services, from pipeline generation to statistical modeling and other data analysis, Shyr said.
"Vanderbilit has a very, very strong precision medicine initiative," Shyr said. Notably, in 2016 Josh Denny, associate professor of biomedical informatics and medicine, got a $71 million budget to build the All of Us research program on behalf of the nationwide Precision Medicine Initiative.
Additionally, the six-year-old Center for Quantitative Computing provides bioinformatics data analysis for the US Centers for Disease Control and Prevention as well for as other universities and some pharmaceutical companies, Shyr said.
The center has two models: outsiders can collaborate with Shyr as a co-investigator or they can hire VANGARD staff on a fee-for-service basis.
Shyr's Department of Biostatistics supports anything involving sequencing data. "Even for some new platforms, we will help investigators develop a new pipeline for that particular platform," he explained. "We have the bandwidth to create any type of omics data."
In bioinformatics, the university-wide working group will consider both genomic and clinical data. Vanderbilt has plenty of both.
In 2015, the university recruited Nancy Cox away from the University of Chicago to run the Division of Genetic Medicine within VUMC and to be the inaugural director of the Vanderbilt Genetics Institute.
At Chicago, Cox and colleagues developed a catalog of genes by medical phenome by applying their previously published PrediXcan analytical approach to more than a decade's worth of EHR data from some 215,000 people. This enabled them to predict the expression of certain genes in 44 different tissues to identify other conditions that might later arise in patients and potential therapeutic approaches.
At Vanderbilt, her division includes BioVU, a DNA bank with more than 240,000 samples. As part of All of Us, Vanderbilt is performing deep genome interrogation on at least 100,000 of those samples. Internally, VUMC is able to cross-reference those results with deidentified clinical data from the EHR, according to Cox.
"It's an astounding opportunity to do science in a different way," Cox said of the advent of the Data Science Visions Working Group. "We're taking a whole variety of approaches that will help us discover things to help us take care of patients better."
For example, she noted that it is not uncommon for patients with high white blood cell counts to have blood tests ordered multiple times. "But some people have a genetic predisposition to high white blood cell counts," Cox noted. Clinicians would be able to reduce test duplication in those patients and spend less time meandering down paths that will lead to dead ends in terms of underlying health issues.
"We need to figure out which parts of genomic data are relevant," Cox said. "We have to educate so many more people in data science."