E-Locus - Institutional Repository of the University of Crete

Home Search

Results - Details

Search command : Author="Μαυροθαλασσίτης" And Author="Γεώργιος"

Current Record: 15 of 26

[Add to Basket]

Identifier

000383048

Title

BioTextQuest:Ένα διαδικτυακό εργαλείο εξόρυξης δεδομένων με στόχο την ανακάλυψη καινούργιας πληροφορίας.Κοινωνικοποίηση γονιδίων:Mια μελέτη της γονιδιακής θέσης ,της περιεκτικότητας σε GC και της σίγασης γονιδίων στη Salmonellα

Alternative Title

BiotextQuest:a data mining tool for concept discovery.Gene socialization :gene order GC content and gene silencing in salmonella

Κοινωνικοποίηση γονιδίων:Μια μελέτη της γονιδιακής θέσης ,της περιεκτικότητας σε GC και της σίγασης γονιδίων στη Salmonella

Author

Παπανικολάου, Νικόλας

Thesis advisor

Ηλιόπουλος, Ιωάννης

Reviewer

Σαββάκης, Χαράλαμπος
Μαυροθαλασσίτης, Γεώργιος
Καραγωγέος Δόμνα
Ηλιόπουλος, Αριστείδης
Πoίράζη, Παναγιώτα
Προμπονάς, Βασίλης

Abstract

This thesis describes research carried out at the Medical School of the University of Crete under the supervision of Professor Charalambos Savakis and in collaboration with Dr Ioannis Iliopoulos. The thesis comprises of 2 distinct parts. The first part describes a text mining method that groups PubMed abstracts in meaningful clusters and the second part describes a whole genome comparison analysis between two bacterial genomes. Part 1: bioTextQuest bioTextQuest is an online tool that allows the user to perform a specialized keyword search in PubMed. The abstracts (that are locally stored in the bioTextQuest Database) are collected and analyzed. The analysis is performed in the following stages: 1. Various predefined words (stoplist) are excluded from the abstracts. 2. Each word of each abstract is weighted for its importance (based on a dictionary) using a variation of a specialized weight algorithm called TF.IDF. Less ‘important’ terms are pruned. Terms with high TF.IDF and terms not appearing in the dictionary pass through. 3. Remaining terms comprise the Li.S.T. (List of Significant Terms). 4. Based on Li.S.T., each abstract is represented by a vector. 5. Various clustering algorithms act on the vectors and group them in clusters. 6. Each cluster is annotated using Gene Ontology (molecular function, cellular compartment and biological process annotation) and Reflect (protein annotation). 7. Each Cluster is presented to users using the respective Significant Terms in a Tag Cloud format that represents the contribution of each term in the corresponding cluster. The clusters can be altered by adjusting several parameters and can be better studied through the aid of their functional enrichment. Clustering can help in quickly assessing a scientific field, concept discovery etc. Part 2: Gene Socialization We performed a genome-wide comparison of two bacterial genomes (Salmonella Typhimurium and Escherichia Coli) focusing on gene order conservation. We study synteny in conjunction with GC content, gene duplication, gene essentiality, gene silencing, horizontal gene transfer and synonymous vs. non-synonymous single-point mutations. We found out that genes that conserve their gene order tend to be more conserved, have higher GC content and lower nonsynonymous/ synonymous ratio. Genes that lose their original position tend to be silenced. Also, duplicated genes follow different evolutionary paths depending on whether they conserve their original position or not: duplicates that remain in their original position tend to be more conserved than the ones that leave their genomic neighborhood. The latter tend to accumulate more AT mutations. Additionally, essential genes tend to remain in their original genetic location.

Language

Greek

Subject

Bioinformatics

Data intergation

Text mining

Whole genome analysis

Βιοπληροφορική

Γονιδιωματική ανάλυση

Εξόρυξη κειμένου

Συγχώνευση γνώσης

Issue date