Date

2018-06-27

Author

Ağyüz, Umut
Purutçuoğlu Gazi, Vilda

Metadata

Show full item record

Item Usage Stats

262
views

0
downloads

Cite This

The breast and gynecological cancers are two most common fatal cancers’ types in women in the world [1]. In oncological literature, these two cancers types are typically worked together since they are the risk factors of each other if the patient has one of these diseases [2]. In general, the cancers, like the heart diseases, are the systems’ ilnesses in the sense that any malfunctions in the associated transaction pathways cause problems in the activation flow, resulting in tumors. Therefore, the mathematical representation of these complex structures enables us to better understand the actual biological process and to produce the target drug. Accordingly, in this study, we evaluate 10 different publically available datasets which are collected from the GEO database. Originally these data are the gene expression datasets where some of them also have certain descriptive information about the samples. Hereby, from each dataset, we initially generate subnetworks by selecting the most significantly expressed genes and normalize them via the RMA method [3] if they are the Affymetrix data or normalize them via deterministic background and quantile normalizations. Then, we present them via distinct mathematical models such as MARS [4] and CGGM [5] in order to describe the steady-state behaviour of the proteomic activations. Here, we evaluate the performance of every model via the accuracy of the estimates and the computational demand. Later, we also combine these models with the risk factors of each cancer and re-construct more comprehensive mathematical models. Finally, we validate our estimated systems via the associated literature and biologically discuss our new findings. By this way, we can also combine the knowledge of both cancers types under a single and more comprehensive mathematical model. We consider that our mathematical representation can open new avenue about these diseases and help us to ask biologically more interesting questions

Subject Keywords

Mathematical models, Breast cancer, Gynecological cancer, Biological systems

URI

https://hdl.handle.net/11511/79391

Collections

Department of Statistics, Conference / Seminar