Ιδρυματικό Αποθετήριο
Πολυτεχνείο Κρήτης
Πολυτεχνείο Κρήτης
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/0F27F7F1-49DD-4E85-9685-50BE30BD4AC1 | ||
| Έτος | 2020 | ||
| Τύπος | Δημοσίευση σε Συνέδριο | ||
| Άδεια Χρήσης |
|
||
| Βιβλιογραφική Αναφορά | I. Torakis, E. S. Bei, S. Sfakianakis, I. S. Pateras and M. Zervakis, “Pancreatic cancer and its correlation with embryogenesis: identification of biomolecular markers using machine learning methods,” in 8th European Medical and Biological Engineering Conference, IFMBE Proceedings, T. Jarm, A. Cvetkoska, S. Mahnič-Kalamiza, D. Miklavcic, Eds., Cham, Switzerland: Springer Nature, 2021, pp. 952–961, doi: 10.1007/978-3-030-64610-3_106. https://doi.org/10.1007/978-3-030-64610-3_106 | ||
| Εμφανίζεται στις Συλλογές |
Pancreatic cancer is a highly lethal disease, projecting to be the second leading cause of cancer-associated deaths. It is considered as one of the most aggressive types of cancer, with one of the major problems reported being the lack of early detection. A patient is diagnosed with pancreatic cancer only in advanced stages, when the possibility of developing a metastasis is high. There is no standard procedure to diagnose high risk patients, since they remain asymptomatic in the cancer’s early stages. Based on the accumulated evidence revealing remarkable parallels in key biological signaling pathways that govern embryonic development and cancer, we sought to extract significant genes at the intersection of these two processes, aiming to identify new tumor markers for pancreatic cancer. Specifically, the aim of this work is to apply machine learning methods to identify biomolecular markers that are differentially expressed in pancreatic cancer patients and correlate them with markers from embryogenesis. After extracting such markers, we use them as predictors within different machine learning methods. Our work contributes a “25 gene signature” of biomolecular markers, which are involved in signaling pathways found in both embryogenesis and pancreatic carcinogenesis, obtained via feature extraction and feature selection methods. These markers are used in classifiers for pancreatic cancer classification and two machine learning models are tested, with good results. We finally justify the notion that our “25 gene signature” can play a classification role in discriminating patients with pancreatic cancer from healthy controls.
Copyright © DIAS 2013
