Incorporating Graph Features for Predicting Protein-Protein Interactions

Incorporating Graph Features for Predicting Protein-Protein Interactions

Martin S.R. Paradesi (Kansas State University, USA), Doina Caragea (Kansas State University, USA), and William H. Hsu (Kansas State University, USA)
Copyright: © 2009 |Pages: 19
DOI: 10.4018/978-1-60566-398-2.ch004
OnDemand PDF Download:
No Current Special Offers


This chapter presents applications of machine learning to predicting protein-protein interactions (PPI) in Saccharomyces cerevisiae. Several supervised inductive learning methods have been developed that treat this task as a classification problem over candidate links in a PPI network – a graph whose nodes represent proteins and whose arcs represent interactions. Most such methods use feature extraction from protein sequences (e.g., amino acid composition) or associated with protein sequences directly (e.g., GO annotation). Others use relational and structural features extracted from the PPI network, along with the features related to the protein sequence. Topological features of nodes and node pairs can be extracted directly from the underlying graph. This chapter presents two approaches from the literature (Qi et al., 2006; Licamele & Getoor, 2006) that construct features on the basis of background knowledge, an approach that extracts purely topological graph features (Paradesi et al., 2007), and one that combines knowledge-based and topological features (Paradesi, 2008). Specific graph features that help in predicting protein interactions are reviewed. This study uses two previously published datasets (Chen & Liu, 2005; Qi et al., 2006) and a third dataset (Paradesi, 2008) that was created by combining and augmenting three existing PPI databases. The chapter includes a comparative study of the impact of each type of feature (topological, protein sequence-based, etc.) on the sensitivity and specificity of classifiers trained using specific types of features. The results indicate gains in the area under the sensitivity-specificity curve for certain algorithms when topological graph features are combined with other biological features such as protein sequence-based features.
Chapter Preview


Protein-Protein Interaction Prediction Problem

The term protein-protein interaction (PPI) refers to associations between proteins as manifested through biochemical processes such as formation of structures, signal transduction, transport, and phosphorylation. PPI plays an important role in the study of biological processes. Many PPIs have been discovered over the years and several databases have been created to store the information about these interactions such as BIND (Bader et al., 2003), DIP (Salwinski et al., 2004), MIPS (Mewes et al., 2002), IntAct (Kerrien et al., 2007) and MINT (Chatr-aryamontri et al., 2007). In particular, more than 80,000 interactions between yeast proteins are available from various high-throughput interaction detection methods (von Mering et al., 2002). These methods can detect if the interaction is either a physical binding between proteins or a functional association between proteins. Often, the functional association between two proteins leads to physical binding among them. Determining PPI using high-throughput methods is expensive and time-consuming. Furthermore, a high number of false positives and false negatives can be generated. Therefore, there is a need for computational approaches that can help in the process of identifying real protein-protein interactions.

Several methods have been designed to address the task of predicting protein-protein interactions using machine learning. Most of them use features from protein sequences (e.g., amino acids composition) or associated with protein sequences directly (e.g., GO annotation). However, the PPI network can be used to design node and topological features from the associated graph. Several methods use such relational and structural features extracted from the PPI network, along with the features related to the protein sequence. This chapter provides an overview of several machine learning methods for predicting PPI using the graph information extracted from a PPI network along with other available biological features of the proteins and their interactions, and shows the importance of the graph features for accurate predictions.

Complete Chapter List

Search this Book: