National Technical University of Athens (NTUA)
Πρόβλεψη της δευτεροταγούς δομής ακολουθιών RNA με έμφαση στο μοτίβο ψευδοκόμβων, με τη χρήση τεχνικών συντακτικής αναγνώρισης προτύπων
Abstract
dc:descriptionAccurate "base pairing" in RNA molecules, which is an essential task in order to predict RNA secondary structures, plays a crucial role in explaining unknown biological processes. The COVID-19 pandemic, a widespread disease, has had a devastating impact on humanity in recent years. This extreme condition enhances the significance of analyzing RNA molecules, and their structures have been highlighted by SARS-CoV-2, a single-stranded RNA virus that causes COVID-19 disease. This thesis aims to develop a pioneering set of frameworks that leverage syntactic pattern recognition to predict specific RNA structures. Specifically, it focuses on introducing novel systems for predicting RNA secondary structure patterns including pseudoknots, utilizing syntactic pattern-recognition strategies, and concepts such as maximum base pairing and minimum free energy. By primarily treating pseudoknot predictions as parsing and optimization problems, the proposed methodologies formalize the prediction of RNA secondary structures. The first developed framework, called Knotify, addresses the prediction of first-order pseudoknots (H-type). It introduces a context-free grammar (CFG) capable of recognizing potential pseudoknot patterns. Knotify exhibits prediction capabilities similar to state-of-the-art frameworks but significantly outperforms them in terms of execution time efficiency. An optimized version of Knotify is also presented, utilizing pruning techniques for further efficiency gains. To predict more complex patterns of H-type pseudoknots, Knotify+ introduced, which extends the power of CFG by searching for intricate patterns in the pseudoknot’s loops such as bulges and internal loops, while incorporating the advantages of maximum base pairing and minimum free energy. Knotify+ demonstrates superior accuracy in predicting core stems compared to existing frameworks. It performs exceptionally well in small sequences and maintains a comparable accuracy rate in larger ones, all while requiring shorter execution times than well-known platforms. Finally, a new grammar-based system is presented to address the prediction of a rare but complex type of pseudoknot, the L-type pseudoknot, that proves effective in prediction and execution time when compared to other novel systems. These innovative systems and architectures aim to improve the ability of biologists to predict RNA motifs, including pseudoknots. They hold the potential for application in various biological domains, such as gene therapy, drug design, and understanding RNA functionality. Moreover, these approaches can be combined with other methodologies to improve the precision of RNA structure prediction. To support this initiative, the Knotify, Knotify+, L-type prediction framework source codes and implementation details are available online as public repositories on GitHub.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Makris, Evangelos
- Μακρής, Ευάγγελος
Subjects
dc:subject × 17- Η-τύπου ψευδοκόμβοι
- L-τύπου ψευδοκόμβοι
- RNA
- Συντακτικοί αναλυτές
- Γραμματικές χωρίς συμφραζόμενα
- Συντακτική αναγνώριση προτύπων
- H-type pseudoknot structure
- L-type pseudoknot structure
- Parsers
- CFG
- Syntactic pattern recognition
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Επιστήμες Μηχανικού και Τεχνολογία
- Υπολογιστές, Υλικό (hardware) και Αρχιτεκτονική
- Electrical Engineering, Electronic Engineering, Information Engineering
- Engineering and Technology
- Computer science, Hardware and Architecture
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/56974