ePrints.FRI - University of Ljubljana, Faculty of Computer and Information Science

Artificial intelligence methods for discovery of relationships in genetic data

Peter Juvan (2005) Artificial intelligence methods for discovery of relationships in genetic data. PhD thesis.

[img] PDF (Disertacija)
Download (953Kb)


    This dissertation reports on research and design of artificial intelligence and bioinformatics approaches and their application in the field of functional genomics. A novel approach to construction of genetic networks from data on mutations is proposed. Network construction involves two steps. The first step, inference of relations between genes, is characterized as abductive. Genetic experiments are the observations that need to be explained and relations between genes are abduced in order to provide explanation of the experimental observations. The second step involves integration of relations into a qualitative network. Several algorithms for construction and interpretation of qualitative networks are presented. Such a network enables qualitative reasoning about the underlying genetic mechanisms. The dissertation also describes a computer program called GenePath, a practical implementation of the approaches proposed in this dissertation. GenePath allows online analysis of genetic data on mutations, emphasizing the importance of exploratory data analysis, transparency of results and machine-based explanation. Additionally, the dissertation analyzes the possibility of constructing genetic networks from quantitative data on gene expression measurements using DNA microarray technology. A previously proposed distance-based approach to inference of relations between genes is described and augmented with a method for estimating their significance. Two alternative statistical approaches are proposed which can handle noise and missing data. The proposed approaches are experimentally evaluated. The utility of GenePath is demonstrated on a number of well-studied genetic analysis problems from D.discoideum and C.elegans. The approaches that deal with quantitative data are applied to study D.discoideum strains with single and double knock-out mutations in genes yakA, pkaC, pkaR, pufA and regA.

    Item Type: Thesis (PhD thesis)
    Language of Content: Slovenian
    Mentor / Comentors:
    Name and SurnameIDFunction
    izr. prof. dr. Blaž Zupan106Mentor
    Link to COBISS: http://www.cobiss.si/scripts/cobiss?command=search&base=50070&select=(ID=5114196)
    Institution: University of Ljubljana
    Department: Faculty of Computer and Information Science
    Item ID: 767
    Date Deposited: 11 Dec 2008 16:09
    Last Modified: 13 Aug 2011 00:34
    URI: http://eprints.fri.uni-lj.si/id/eprint/767

    Actions (login required)

    View Item