The team of researchers are examining ways to quickly determine side effects from medications using machine learning methods. They use medical informatics resources and a human genome-scale metabolic model to present the array of model-based phenotype predictors (AMPP) to identify these probems which can include intersitial nephritis and other pyramid disorders. Drug side affects "levy a massive cost on society through drug failers, morbidity, and mortality cases every year..[.]" The group has determined that AMPP predicts, rather substantially, (AUC > 0.7) for >70 drug side effects. AMPP also is a better indicator than a previous biochemical structure-based method for the prediction of metabolic side effects.
This work is significant as it will help to reduce side effects that are metabolic in nature--particularly as scientists develop new drug treatments for various diseases.
The Department welcomes comments, suggestions and corrections. Send email to editor [at] cs [dot] umd [dot] edu.