Mining disease fingerprints from within genetic pathways.
Year: 2012
Type: article
Abstract: Mining biological networks can be an effective means to uncover system level knowledge out of micro level associations, such as encapsulated in genetic pathways. Analysis of human disease genetic pathways can lead to the identification of major mechanisms that may underlie disorders at an abstract functional level. The focus of this study was to develop an approach for structural pattern analysis and classification of genetic pathways of diseases. A probabilistic model was developed to capture characteristic components ('fingerprints') of functionally annotated pathways. A probability estimation procedure of this model searched for fingerprints in each disease pathway while improving probability estimates of model parameters. The approach was evaluated on data from the Kyoto Encyclopedia of Genes and Genomes (consisting of 56 pathways across seven disease categories). Based on the achieved average classification accuracy of up to ~77%, the findings suggest that these fingerprints may be used for classification and discovery of genetic pathways.
Source: PubMed Central
Authors: Ahmed Ragab Nabhan, Indra Neil Sarkar
Institution: University of Vermont
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W167856683