New insights achieved by the combination of bioinformatics and genomics data through machine learning could help the way doctors treat chronic diseases like lupus, according to a paper published in the journal Nature Scientific Reports.
George Washington University researchers, including Keith A. Crandall, PhD, director of the Computational Biology Institute at Milken Institute School of Public Health (Milken Institute SPH), worked with AMPEL BioSolutions, a genomics technology company, to train a computer to analyze genetic and patient data to predict whether an individual living with lupus was experiencing a flare in disease activity. The genetic analysis, called gene expression analysis, examines the number and pattern of genes expressed at a single moment and can provide insight into various genomic abnormalities.
This innovative machine learning approach can be used for many autoimmune or inflammatory diseases, which can have unexpected flares of activity that affect a patient’s quality of life. Physicians one day may be able to use information gathered by a lab test and analyzed by this approach to predict flares and provide early treatment, saving patients from unnecessary pain and other symptoms.
“This work is a great example of the insights gained by using artificial intelligence such as machine learning approaches to analyze diverse types of data, in this case integrating clinical data with genomics, to make effective and individualized insights on patient health,” said Crandall, who is also a professor of biostatistics and bioinformatics at Milken Institute SPH.
Next steps of the project include expanding the study to more individuals and more genomic data types, and then developing the physical lab test for physicians to use and conducting clinical trials. The paper, “machine learning approaches to predict lupus disease activity from gene expression data”, was published July 3 in Nature Scientific Reports.