Abstract
The analysis of structure and dynamics of biological networks plays a central role in understanding the intrinsic complexity of biological systems. Biological networks have been considered a suitable formalism to extend evolutionary and comparative biology. In this paper we present GASOLINE, an algorithm for multiple local network alignment based on statistical iterative sampling in connection to a greedy strategy. GASOLINE overcomes the limits of current approaches by producing biologically significant alignments within a feasible running time, even for very large input instances. The method has been extensively tested on a database of real and synthetic biological networks. A comprehensive comparison with state-of-the art algorithms clearly shows that GASOLINE yields the best results in terms of both reliability of alignments and running time on real biological networks and results comparable in terms of quality of alignments on synthetic networks. GASOLINE has been developed in Java, and is available, along with all the computed alignments, at the following URL: http://ferrolab.dmi.unict.it/gasoline/gasoline.html.
Highlights
The structure and the dynamic of biological networks arise from interactions among molecules within the cell
We present GASOLINE (Greedy And Stochastic algorithm for Optimal Local multiple alignment of Interaction NEtworks), a novel algorithm for protein networks alignment based on iterative sampling [26] in connection with a greedy strategy
Given N weighted biological networks, where weights are probabilities expressing the reliability of pairwise protein relations, informally, the local alignment of biological networks aims at finding a set of N regions, one from each network, that are conserved in their sequence and interaction pattern
Summary
The structure and the dynamic of biological networks arise from interactions among molecules within the cell. A retrospective view of the recent history of molecular biology research shows that most of the attention has been devoted to sequence analysis. This represents a fundamental level of biological investigation and for a long time has been the basis of evolutionary studies [12]. Local network alignment approaches based on Hidden Markov Models have been proposed [24,25]. Their action has been restricted to the identification of shared paths
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
