New Data Tool Enhances Drug Safety, Aids Search for New Uses of Existing Drugs
Sarangdhar M, Tabar S, Schmidt C, Kushwaha A, Shah K, Dahlquist JE, Jegga AG, Aronow BJ. Data mining differential clinical outcomes associated with drug regimens using adverse event reporting data. Nat Biotechnol. 2016 Jul 12;34(7):697-700.
Mayur Sarangdhar, PhD, under the direction of Bruce Aronow, PhD, and Anil Jegga, DVM, developed a new web-based analytical tool called AERSMine that allows any website visitor to carry out sophisticated clinical drug response outcome analyses across millions of patients using a powerful new database approach pioneered in the Division of Biomedical Informatics. Anyone from physicians, to researchers, to the general public can use the open-source tool to rapidly find, combine, and analyze the growing volume of drug information stored in the U.S. Food and Drug Administration’s Adverse Event Reporting System (FAERS). AERSMine allows users to create virtual cohorts and identify patterns of differential short- and long-term outcomes as a function of medication, underlying disorders, other co-occurring conditions, age and gender. The database also has the potential to help researchers develop novel treatments for diseases by identifying improved uses of individual drugs or drug combinations.
New Bioinformatics Pipeline Helps Interpret Fate of Blood Cells
Olsson A, Venkatasubramanian M, Chaudhri VK, Aronow BJ, Salomonis N, Singh H, Grimes HL. Single-cell analysis of mixed-lineage states leading to a binary cell fate choice. Nature. 2016 Sep 29;537(7622):698-702.
Nathan Salomonis, PhD, developed a new bioinformatics computer program as part of his AltAnalyze suite called Iterative Clustering and Gene-Guided Selection (ICGS) to give researchers a comprehensive view of the overwhelming amount of data generated by single-cell RNA sequencing. ICGS offers an intuitive platform that can comprehensively process and analyze sequencing as well as biological data to identify various transitioning or shifting genomic and cellular states of cells. Salomonis worked with H. Leighton Grimes, PhD, and his team, to blend laboratory biology with this new bioinformatics approach to better understand how blood cells develop. The researchers used single-cell RNA sequencing technology to identify and study the different genes and regulatory networks within individual cells. Identifying the shifting genomic states of cells is possible by using the ICGS pipeline. Researchers found developing blood cells caught in tugs of war between competing gene regulatory networks before finally deciding what type of cell to become. The research offers a foundation for better understanding how blood and immune system disorders develop.
Machine Learning Automates Identification of Verbal, Nonverbal Suicidal Behaviors
Pestian JP, Sorter M, Connolly B, Cohen KB, McCullumsmith C, Gee JT, Morency L-P, Scherer S, Rohlfs L. A Machine Learning Approach to Identifying the Thought Markers of Suicidal Subjects: A Prospective Multicenter Trial. Suicide Life Threat Behav. 2017 Feb;47(1):112-121.
John Pestian, PhD, MBA, and his team are developing new ways to predict risk of suicide. Suicide is the third leading cause of death among 15-25 year olds. Determining whether or not a patient is suicidal can be a major challenge, and one that was largely unsupported by technology, until now. Pestian and his team have created a new machine learning algorithm that can analyze verbal and nonverbal cues to help identify suicidal individuals. In a study, they used the algorithm to analyze subjects’ words and vocal characteristics—such as tone, pauses, and pitch—in order to classify their behavior as suicidal, mentally ill but not suicidal, or neither. Results showed that the machine learning algorithm could classify subjects with up to 93 percent accuracy. The study provides insight into how to use advanced technology as a decision support tool to identify and prevent suicidal behavior.
