Not Logged In

Predicting Homologous Signaling Pathways Using Machine Learning

Full Text: Bostan_Babak_Fall 2009.pdf PDF

Understanding biochemical reactions inside cells of individual organisms is a key factor for improving our biological knowledge. Signaling pathways provide a road map for a wide range of these chemical reactions that convert one signal or stimulus into another. In general, each signaling pathway in a cell involves many different proteins, each with one or more specific roles that help to amplify a relatively small stimulus into an effective response. Since proteins are essential components of a cell’s activities, it is important to understand how they work – and in particular, to determine which of specie’s proteins participate in each role. Experimentally determining this mapping of proteins to roles is difficult and time consuming. Fortunately, many individual pathways have been annotated for some species, and the pathways of other species can often be inferred using protein homology and the protein properties.
We present an automatic approach, PSP, that uses the signaling pathways in well-studied species to predict which proteins will serve which roles in less-studied species. Our machine learning approach creates a predictor that achieves a generalization F-measure of 78.2% when predicting protein roles in 11 different pathways across 14 different species. We also describe an evaluation method that suggests our prediction might be more accurate than this F-measure. This method makes predictions based on historical data, then evaluates the prediction based on new data that include more recent annotations of the proteins. This process revealed that our historical predictor was correct about many predictions that were considered to be wrong based on the historical data.

Citation

B. Bostan. "Predicting Homologous Signaling Pathways Using Machine Learning". MSc Thesis, October 2009.

Keywords: Signaling Pathway, Machine Learning, Protein prediction
Category: MSc Thesis

BibTeX

@mastersthesis{Bostan:09,
  author = {Babak Bostan},
  title = {Predicting Homologous Signaling Pathways Using Machine Learning},
  year = 2009,
}

Last Updated: October 13, 2009
Submitted by Babak Bostan

University of Alberta Logo AICML Logo