| Description |
-
Abstract: Cross-species biological network alignments are powerful tools for knowledge discovery and have been a major focus in the field of computational biology. Many alignment algorithms have been published in recent years, along with validation techniques to quantify the success of an alignment. In this work we show that existing validation techniques are not always suitable metrics to assess ... read morean alignment's performance. We evaluate the validation techniques using a collection of network perturbation tests, and extend our analyses to several cross-species alignments. This thesis begins by introducing the field of biological network alignment, including the protein protein interaction (PPI) networks, existing alignment algorithms, and existing validation techniques. The focus then shifts towards a description of the experiments performed in this work, followed by an analysis of the results that reveals the inadequacies of existing validation techniques.
Thesis (M.S.)--Tufts University, 2015.
Submitted to the Dept. of Computer Science.
Advisor: Benjamin Hescott.
Committee: Donna Slonim, and Mark Crovella.
Keyword: Computer science.read less
|
|---|
This object is in collection
