With the recent publication of the first paper to come out of the University of Chicago’s collaboration with Google, our director Sam Volchenboum sat down with The Forefront for a Q&A about the project. Read the interview to learn more about how the researchers leveraged EHR data from the CRDW for the development of new predictive algorithms for heathcare, how patient privacy is protected throughout the process, and where Sam thinks the future of EHR data mining could take us.
Machine learning algorithms are getting better and better at predicting and diagnosing disease — even though sometimes, researchers can’t fully explain why. In a recent paper supported by CRI effort and CRDW data, researchers from the University of Chicago, Stanford University, the University of California, San Francisco, and Google determined that these algorithms could use patterns in patient data to predict diseases and the likelihood of certain medical outcomes with extraordinary accuracy.
In the Harvard Business Review, CRI Director Sam Volchenboum and Immuta Chief Privacy Officer and Legal Engineer Andrew Burt write, “This future is alarming, no doubt, due to the power that doctors and patients will start handing off to machines. But it’s also a future that we must prepare for — and embrace — because of the impact these new methods will have and the lives we can potentially save.”
In a paper published in the May 4, 2018 issue of Nature Communications, University of Chicago researchers confirmed the existence of a curable intermediate state of cancer called oligometastasis. Samuel Hellman, MD, Ralph Weichselbaum, MD, and their co-authors were able to use molecular analysis to reliably predict which patients with colorectal cancer could benefit from surgery. Their findings should improve the treatment of patients with colorectal metastases that have spread to the liver, and could potentially be applicable to many types of cancer.
The CRI contributed bioinformatics analysis to the study, with our Senior Bioinformatician Lei Huang, PhD, and Director of Bioinformatics and Research Assistant Professor Jorge Andrade, PhD, named as co-authors.
Read more: Study confirms curable state between single and widespread cancers
A University of Chicago team including the CRI’s Riyue Bao has been honored with a Team Science Award from the Melanoma Research Alliance. The Team Science Awards provide funding for multidisciplinary teams to carry out high-impact melanoma research with the potential to rapidly lead to clinical advances.
The PIs on the UChicago team are Thomas Gajewski, MD, PhD; Jason Luke, MD; and Cathryn Nagler, PhD, with our own Research Assistant Professor and Manager of Bioinformatics Riyue Bao, PhD, named as Young Investigator. The award will support their continued work in investigating how the composition of patients’ gut microbiota mediates the efficacy of anti-PD-1 immunotherapy treatments for melanoma.
Using new gene editing tools, UChicago researchers have identified key genes that play a role in whether the flu virus can infiltrate a lung cell. For their study, which was published in the April 10, 2018 issue of Cell Reports, Julianna Han and Balaji Manicassamy, PhD, used CRISPR/Cas9 gene editing tools to create a library of cells, each missing a different gene. They then exposed each of these cells to the H5N1 flu virus. After selecting for cells that survived the exposure, they partnered with the CRI Bionformatics Core to analyze which genes could have played a role in the cells’ resistance. This analysis identified two key genes involved in the H5N1 response.
Through research like this, scientists are developing a better understanding of what proteins and pathways play a role in flu infections, which should help with the development of the next wave of targeted antiviral treatments. The CRI’s Yan Li, PhD, and Jorge Andrade, PhD, were co-authors on this study.
In an interview with Technology Networks, CRI Director Sam Volchenboum offers insights into the future of using big data to fight pediatric cancer and other diseases. In particular, he highlights the need for standardized data models, which will allow researchers to harness the true power of big data by making it easier for data from disparate sources to be harmonized and combined. In addition, advances in genomic research will be made possible when genomic data is linked with more and richer sources of clinical data. The CRI is at work on both of these issues with our initiative in pediatric cancer research data commons.
The CRI’s development of an enterprise Clinical Trials Management System (CTMS) to unify, standardize, and modernize clinical research efforts across the BSD continues on track this spring. Since the source code development phase began in August 2017, our team has been working to develop the key components of the system, and are currently on schedule with this development process. The team’s Continuous Integration development method has each iteration of the software in the hands of testers and stakeholders early and often and allows early detection of any potential code integration issues.
In October, the Center for Research Informatics and the Graham School MSc in Biomedical Informatics program jointly launched a new monthly speaker series, the Computational Life Science Seminar Series (CLSSS). The CLSSS was created to bring together researchers from all life sciences fields and offer a forum to highlight new research and foster collaborations. The format is flexible and sessions are open to faculty, students, and researchers throughout the University of Chicago community.
Thus far, the series has hosted speakers from the Center for Data Intensive Science to showcase data commons initiatives, as well as from the Research Computing Center to present their XROMM data management system (read a recap of this session here). The next seminar, scheduled for January 25, will present research currently underway in Jack Gilbert’s laboratory looking at how genomics of the microbiome might be connected to health and obesity outcomes. Register here, and join the mailing list to stay informed about future CLSSS sessions.
In a paper published in the January 5, 2018 issue of Science, a team of University of Chicago researchers led by Thomas Gajewski, MD, PhD, demonstrated a strong link between certain strains of human gut bacteria and success rates for immunotherapy treatments for advanced melanoma.
The CRI’s Manager of Bioinformatics Riyue Bao, PhD, played a key role as coauthor of this paper. Riyue conducted bioinformatics analysis of microbiome and genomic data to explore the differences in the microbial environments of patients who did and did not respond to treatment, as well as integrating multiple types of microbial and genomics data to identify molecular alterations associated with clinical outcome. She also conducted statistical analysis of clinical data used in the study, developed new approaches for interpretation of the analysis results, and contributed to writing the paper.
The results of this study have the potential to improve cancer treatment by predicting which patients will respond well to certain therapies, as well as opening the door to probiotic treatments that could enhance the effects of immunotherapy drugs.
The Graham School MSc in Biomedical Informatics program has partnered with online education company GetSmarter to offer an 8-week online course in Healthcare Informatics. CRI Director Sam Volchenboum serves as Course Convener, guiding the course design and teaching some of the modules alongside other University experts. The online course was developed in order to bring the University of Chicago’s expertise in healthcare informatics to a worldwide audience, preparing students to enter a field that is becoming more complex and demanding by the year as more and more healthcare data is made available to study. Learn more about why we’re offering this course here. The next session begins February 28.
