• Login
    View Item 
    •   Repository Home
    • Electronic Theses and Dissertations
    • Masters Theses and Dissertations
    • Department of Biological Sciences
    • View Item
    •   Repository Home
    • Electronic Theses and Dissertations
    • Masters Theses and Dissertations
    • Department of Biological Sciences
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    APPLICATION OF MACHINE LEARNING IN IDENTIFYING ASSOCIATIONS BETWEEN CONGENITAL INFECTIONS AND IMMUNE DEVELOPMENT IN INFANTS

    Thumbnail
    View/Open
    JOYCE KABAGENYI.pdf (2.848Mb)
    Date
    2020-09-23
    Author
    KABAGENYI, JOYCE
    Metadata
    Show full item record
    Abstract
    The first year of life is a crucial window in the development of an infant’s immune system. Importantly, vaccination for almost all vaccine preventable infections is primarily given within this time. Congenital infections may impair immune function and development in infants. Although these infections have known clinical indications in new-borns, their effects on vaccine efficacy in infants remains poorly understood. Examining such associations requires that we examine the kinetics of immune responses over the first months of life. The hitch however, is in handling such extensive longitudinal data for which there are currently no systematic methods of analysis that account for data structure. This study set out to develop a novel workflow necessary for mining highly dimensional and multi-structured data using unsupervised machine learning algorithms, which would enable us to explore the effects of congenital infections on immune responses to vaccines in infants. Data on immune response to vaccines and congenital infection status was generated using a recently developed protein microarray assay. Multiple machine learning algorithms were explored to determine ranks of responses to vaccines and infections by infants in the first year of life. Finally, associations between congenital infections and vaccine outcomes were assessed using random forests. A new unbiased approach for analysis of longitudinal multidimensional serological data was developed. The prevalence of congenital infections was low among infants in this cohort except for congenital CMV, HSV1 and Parvovirus. Infants clustered into two groups based on IgG responses to vaccines and common childhood infections. Exposure to CMV and HSV were important in predicting infant responses to infections at 1 year of age. This unsupervised approach provides a novel unbiased way to mine multidimensional data for biological patterns using bioinformatics tools. The biological results point to possible dampening of immune response by congenital CMV and HSV infections. These identified congenital exposures may have a significant effect on serological responsiveness of infants later in life and consequently affect vaccine efficacy and immune development in infants. The data generated in this study vi can be leveraged by immunologists to further investigate the mechanistic basis for immunological attenuation by these infections in early
    URI
    http://elibrary.pu.ac.ke/handle/123456789/888
    Collections
    • Department of Biological Sciences

    DSpace software copyright © 2002-2016  DuraSpace
    Contact Us | Send Feedback
    Theme by 
    Atmire NV
     

     

    Browse

    All of PUSpaceCommunities & CollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

    My Account

    LoginRegister

    DSpace software copyright © 2002-2016  DuraSpace
    Contact Us | Send Feedback
    Theme by 
    Atmire NV