Systems, methods and computer-readable media are provided for facilitating patient health care through automatic discovery, establishment, and statistical validation of diagnostic-signals in repositories of EMR information, identifying phenotypic diagnostic signals regarding markers or combinations or sequences of patient attributes. In an embodiment, a method includes identifying, from a dataset of health care records, a first group of patients that have one or more markers; extracting one or more diagnostic statuses that pertain to the group of patients tested with diagnostics such as whole genome sequencing (WGS) or whole exome sequencing (WES); determining one or more statistically significant differences involving frequent itemsets comprised of two or more marker combinations that occur concomitantly in members of the group; determining statistical disproportionality signals among the items and itemsets with regard to the likelihood that WGS or WES will yield a specific, actionable diagnosis; and determining membership of an individual in a diagnostic-yield subset of the group according to the individual's manifesting one or more of the identified signals.