My Role
Data Scientist
Background
Metastasis is the primary reason for mortality in most cancers, causing 90% of all cancer deaths. Breast cancer is considered to be an incurable disease, and although surgery in combination with general oncological treatment prolongs the survival in metastatic cancer, there are many patients who cannot be helped today. The reason is still unknown, why some tumors metastasis whereas others do not. A deeper understanding of the metastatic process would be most valuable to improve and individualized treatment in this patient group.
MicroRNAs (miRNAs) are small, ~19-22 nucleotides, non-coding single stranded RNA molecules. miRNAs are found in both plants and animals, and are thought to regulate diverse biological processes such as cell proliferation, apoptosis and cell differentiation. miRNA deregulation is correlated to human cancers, with potential roles of tumor suppressors or oncogenes. This subgroup of miRNAs associated with cancer, is often referred to as ‘oncomirs’.
Purpose
To study the differences in miRNA expression for breast cancer between normal tissue, primary tumors and liver metastases, and to evaluate the differences between metastatic and non-metastatic primary tumors, with the aim of finding potential therapeutic targets and contribute to the understanding of the mechanisms behind metastasis.
Data
Microdissected epithelial cells from normal tissue, primary tumor and liver metastases from breast cancer patients have been analyzed with microRNA arrays. The raw data material consists of 381 miRNAs, for three different tissue types, from 11 individuals.
Model Used: Mixture model for zero heavy data and empirical Bayes
Mixture model for zero heavy data and empirical Bayes is used to analyze the inference of miRNAs and cancer. Four types of false discovery rates are being used to prevent type 1 error. These are Benjamini Hochberg, Benjamini Yekutieli, Bonferroni and the Bayesian approach local fdr.
Findings
By analyzing the contrast with the corresponding false discovery rates statistically significant miRNA related to cancer, oncomirs where discovered. The study show 39 miRNA with significant differences in expression between normal tissue and primary tumors, and 37 between normal tissue and liver metastases. Only one miRNA show significant difference between primary tumor and liver metastases. Our findings include indication that the expression increases for miRNA15b in cancer condition. A clinical reduction of miRNA15b could in the future be a potential therapeutic target for cancer.