A machine learning model developed by researchers at the Johns Hopkins Kimmel Cancer Center filters out the biological noise in liquid biopsy samples, helping clinicians better match therapies to their patients’ tumors.