Analytical Description of Gene-Protein-Gene Interaction Using Log linear Model in Breast Cancer Studies

Abstract

This study was built on the premise that better diagnosis of cancers has been associated with early detection. Genetic studies have been identified as a good intervention tool to improve diagnosis and little is known about gene-protein-gene interaction. Therefore this study was designed with the aim of detecting gene-protein-gene interactions. The objective of this study was thus to have a better understanding of the interactions among genes and breast cancer biomarkers using a more robust statistical model.

The study was carried on the premise that breast cancer is the top cancer in women in the developing world and particularly in Africa where it occurs with the highest incidence. Most risk-reduction strategies cannot eliminate the majority of breast cancers that develop in countries of the developing world where breast cancer is diagnosed in very late stages. As a result early detection in order to improve breast cancer outcome and survival remains the cornerstone of breast cancer control. This justified the timing of this study.

A log-linear built model was fitted into breast cancer data available at the Advanced Medical Research and training Institute of the University College hospital-University of Ibadan-Nigeria. Secondary breast cancer data was collected from some breast cancer patients at the Institute of Advanced Medical Research and Training Institute (I.M.R.A.T), of University College Hospital (U.C.H) Ibadan-Nigeria. Different levels of Estrogen receptor alpha (E.R), Prostrogen Receptor (P.R) and HER-2-Neu were analyzed in this study. Data was reduced to contingency tables whose effect was distortion of their continuous nature into discrete form. Although this led to a loss of data identity, it was considered necessary to make computations feasible.

The t-test statistic (analog of the likelihood ratio statistic) and Wald test were used to test the extent of fit (level of statistical significance) of the regression coefficients (the betas).  The findings revealed that there is a statistical significance in the interaction between Prostrogen Receptor (P.R) and HER-2-NEU only when Estrogen receptor alpha (E.R) level was low. This implies that patients with breast cancer will benefit more from treatment at this level which also corresponds to the early stages of breast cancer.

Keywords: Log linear Model, Breast cancer