An Improved Archaeology Algorithm Based on Integrated Multi-Source Biological Information for Yeast Protein Interaction Network

Abstract

With the development of high-throughput interaction detection techniques such as tandem affinity purification (TAP) and yeast two-hybrid (Y2H), the available genome-wide protein–protein interactions (PPIs) data have been increasing in recent years. Using mathematical, physical, and artificial-intelligence methods, some researchers in computational biology focused on uncovering the evolutionary ages of proteins according to present PPI networks (PINs), but improving their accuracy was challenging. A plausible explanation is that they solved biological problems with non-biological techniques and did not provide much attention to biological backgrounds and meanings of proteins or their relationships. In this paper, we propose two ways to improve the accuracy of age predicting and skillfully “embedding” multi-source biological information in each iteration of an archaeology algorithm for yeast PIN. On the one hand, we reduce the probability of reversing errors by decreasing the non-duplication protein pairs, which are obtained from 460 gene trees constructed by means of a multiple sequence alignment and the neighbor joining algorithm. On the other hand, the reliable crossover standard from different biological information sources can decrease local random errors of alternative treatment. The application of the novel algorithm to simulation data and real yeast PINs shows a marked improvement in accuracy. Our research strongly suggests that putting non-biological methods into the “biological context” will bear more favorable results.