Systems and methods for classifier retraining for sub-populations with unique epigenetic patterns

Abstract

Systems and methods of the disclosure may include receiving a first set of nucleic acid sequencing data associated with a general population and a second set of nucleic acid sequencing data associated with a subpopulation, obtaining a first feature set derived from a first set of one or more features based on the first set of nucleic acid sequencing data and a second feature set derived from a second set of one or more features based on the second set of nucleic acid sequencing data; training, by utilizing the first feature set and the second feature set as training input, a machine learning model to predict a disease condition; generating a filtered feature set by filtering at least a subset of features from the second feature set; and retraining the machine learning model to predict the disease condition using the second filtered feature set and the first feature set.

Images