Removal of cell-free DNA from test samples for classification by mixture models

Abstract

A method and system for removing non-indicative test sequences from a test sample, including both indicative and non-indicative test samples, is disclosed. Generally, the non-indicative test sequences identified are derived from WBC cfDNA, while the indicative test sequences are derived from cfDNA associated with identifying the presence of cancer in the sample. To identify the non-indicative test sequences, the system applies a disambiguation model to the test sequences in the test sample. The disambiguation model matches genomic regions from the test sequences in the test sample with genomic regions of test sequences from a sample cohort. The model then generates a feature set for the matched test sequences and determines the probability that the sequences from the test sample represent WBC-derived cfDNA. In representing WBC-derived cfDNA, the system removes the test sequences from the sample to form a classifier and applies a cancer classifier to the classifier.

Images