Abstract: Provided herein are methods and systems for analysing a sample of a subject by using a trained algorithm to classify the samples as invasive, non-invasive, metastatic, non-metastatic or combinations thereof. The methods may comprise using gene signatures as features of a model. The method may comprise using one or more gene aberrations as features of a model.

Abstract: Gene fusions are hybrid genes formed by two previously separate genes. Many such gene fusions may be strong driver mutations for cancer and may play important roles in tumorigenesis. Gene fusions may be identified in biological samples extracted from the diseased location. However, for lung cancer, extracting samples from the nodules may be challenging and may lead to undesirable consequences. Disclosed herein is a method and system of identifying a gene fusion landscape from bronchial brushings and further identifying fusions that may be potentially associated with lung cancer.