Breast Cancer Biomarker Identification via Metabolomics Data

Abstract

Breast cancer (BC) is the most common cancer in women. Although it has a relatively high survival rate, the importance of non-invasive biomarker research is increasing to supplement the limitation of the mammography and early diagnosis. To deal with the constraints of direct human experiments, a canine mammary tumor (CMT) can be a suitable animal model in BC because of the similarity of physiological characteristics. Metabolomics, the study of the metabolic response of the living system, identify the significantly differentiated metabolites in cancer. The purpose of this research is profiling the metabolite biomarker candidates altered in urine samples from CMT for early diagnosis of CMT. All metabolites were quantified using nuclear magnetic resonance (NMR). Moreover, statistical and machine learning methods were used for biomarker candidate selection. As a result, I found a total of 20 metabolites as CMT biomarkers. For more substantial biological interpretation, I performed metabolic pathway analysis with metabolite biomarker candidates. The CMT metabolic biomarkers not only suggest the early diagnostic biomarker candidates but also explain the altered metabolism in canine cancer.