Unit:
Κατεύθυνση ΒιοπληροφορικήΠληροφορική
Author:
Savva Dimitrianos
Supervisors info:
Γεώργιος Σπύρου, Ειδικός Λειτουργικός Επιστήμονας Α ́, Ίδρυμα Ιατροβιολογικών Ερευνών, Ακαδημίας Αθηνών
Original Title:
Υπολογιστικές μέθοδοι βελτιστοποίησης τυχαίων περιπάτων σε γονιδιακά δίκτυα
Translated title:
Computational methods for random walk optimization on gene networks
Summary:
Huge biological networks are being published everyday as a result of high throughput technologies. Because of the huge volume of data that has been made available, many systemic approaches have been developed to highlight hidden information within these biological networks. Among such approaches, clustering and random walks-based techniques have shown hopeful results about the discovery and ranking of important genes and genetic correlations for numerous diseases. However, there is a need for improvement of those approaches because of the increasing size and complexity of the biological networks. The main goal is the reduction of the computing time and cost to allow for deeper exploration of biological networks. In this particular study, we developed the XInfoWalk, a tool for exploring biological networks using informed random walk algorithms. More specifically, the goal of this software is the reduction of the computing time of the Informed Walks algorithms already developed in our group, by employing efficient techniques. The Informed Walks model explores gene networks using biologically informed random walks. It has been used for the discovery and ranking of important genes and genetic correlations for seven types of cancer. XInfoWalk was used in microarray data from patients with myocardial infarction for the discovery of important genes. XInfoWalk is an optimized version of Informed Walks that reduces dramatically the required running time and makes feasible longer runs. A case study for XinfoWalk has been performed using microarray data from patients with myocardial infarction for the discovery of important genes. The discovered gene network signatures, have been used in the investigation of possible repurposed drugs and related molecular mechanisms.
Main subject category:
Science
Keywords:
random walk optimization, biological networks, drug repositioning, biological mechanisms, statistical analysis
