{"id":847,"date":"2022-08-15T15:43:15","date_gmt":"2022-08-15T15:43:15","guid":{"rendered":"http:\/\/new.patterncomputer.com\/?p=847"},"modified":"2022-08-23T18:22:16","modified_gmt":"2022-08-23T18:22:16","slug":"pattern-computer-and-university-of-washington-discover-viral-sequences-predicting-complications-in-conjunctivitis","status":"publish","type":"post","link":"http:\/\/new.patterncomputer.com\/pattern-computer-and-university-of-washington-discover-viral-sequences-predicting-complications-in-conjunctivitis\/","title":{"rendered":"Pattern Computer and University of Washington Discover Viral Sequences Predicting Complications in Conjunctivitis"},"content":{"rendered":"

REDMOND, WA, August 15, 2022<\/strong> — Pattern Computer\u00ae, Inc. (PCI) and researchers led by a team from University of Washington have, for the first time, demonstrated that the development of subepithelial Infiltrates (SEI) due to conjunctivitis infection can be accurately predicted from knowledge of the full viral sequence. Their work published in Ophthalmology Science<\/em> (https:\/\/doi.org\/10.1016\/j.xops.2022.100166<\/a>) suggests that the development of SEI in adenovirus D8 (AdV D8) conjunctivitis is largely attributable to pathologic viral sequence variants within the D8 type and establishes machine learning paradigms as a powerful technique for understanding viral pathogenicity.<\/p>\n

Conjunctivitis is among the most common infectious conditions worldwide. Usually, conjunctivitis leads to mild symptoms such as pink eye, but some infections may result in SEI, which can lead to longer term visual impairment. To understand the degree to which viral sequence variants determine clinical outcome in adenoviral conjunctivitis, Pattern Computer, in collaboration with researchers in the Department of Ophthalmology at the University of Washington and collaborators from Harvard Medical School and Novabay, analyzed 80 samples extracted from patients infected with one of the leading causes of conjunctivitis, Adv D8. In a parallel approach, the University of Washington and Pattern Computer teams used, respectively, extreme random forests and Pattern Computer’s proprietary Pattern Discovery Engine\u2122, to analyze the data. Both approaches generated models that demonstrated 100% accuracy in predicting the country of origin, viral clade (subclass) and the likelihood for SEI development on a holdout test set of 16 samples.<\/p>\n

While predictive power is important, Forests and ensemble models are hard to interpret. The models generated by the Pattern Discovery Engine summarize all the genomic and SEI risk information in a set of simple, interpretable, and actionable equations that, for the first time, relate Adv D8 genomic signatures to the development of SEI; and provide insights into the mutations that may make a viral strain more likely to cause SEI than others.<\/p>\n

According to senior author Russell N. Van Gelder, MD, PhD, the Boyd K. Bucey Memorial Chair, professor and chairman of the Department of Ophthalmology at University of Washington, “This work establishes the power of machine learning models to establish the roots of pathogenicity and provide testable hypotheses of causation for outcomes in infectious disease. The ability of the Pattern Computer Discovery Engine to provide concrete models with high explainability facilitates our next steps in understanding the pathogenesis of viral conjunctivitis. This approach will doubtless be very useful for the study of many other infectious diseases.”<\/p>\n

Although larger datasets are needed to generate universal models of SEI risk, the collaboration between the University of Washington and Pattern Computer, suggests, for the first time, that:<\/p>\n