Construction of a New Model to Investigate Breast Cancer Data

Ağyüz U., PURUTÇUOĞLU GAZİ V., Purutçuoğlu E., ÜRÜN Y.

6th International Conference on Dynamics Games and Science, DGS-VI-2018, 19th Latin American Conference on Economic Theory, JOLATE 2018 and 21st International Consortium on Applied Bioeconomy Research, ICABR 2017, California, United States Of America, 30 May - 02 June 2017, vol.365, pp.11-29 identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 365
  • Doi Number: 10.1007/978-3-030-78163-7_2
  • City: California
  • Country: United States Of America
  • Page Numbers: pp.11-29
  • Keywords: Breast cancer, Copula Gaussian graphical model, Generalized additive models, Multivariate adaptive regression splines model, Systems biology
  • Middle East Technical University Affiliated: Yes


© 2021, Springer Nature Switzerland AG.Modelling is a way to describe the elements of the network/system, their states and their interactions with other elements in order to understand the current state of knowledge of a system. Thereby, the mathematical models may predict the experiments which are difficult or impossible to do in the lab and can be used to discover indirect relationships between model’s components. Hereby, the aim of this study is to develop a network structure for the breast cancer from the analyses of different datasets which include the data of the luminal type at the stage 1–3 breast cancer diagnosed in total 377 patients and related to the PI3KCD signalling pathway. Accordingly, in the analyses, the relations of the 65 oncogenes are revealed by a true network in a binary format. Then, we construct the quasi breast cancer networks by using different parametric and non-parametric models, namely, Gaussian graphical model, copula Gaussian graphical model and multivariate adaptive regression splines with/without interaction terms. In the computations, we evaluate the performance of all suggested mathematical models via F-measure and accuracy measure criteria. We consider that the outcomes can be useful for the selection of the best fitted model in the construction of the breast cancer gene-gene interaction networks.