A New Family of Similarity Measures for Scoring Confidence of Protein Interactions using Gene Ontology

Abstract

The large-scale protein-protein interaction (PPI) data has the potential to play a significant role in the endeavor of understanding cellular processes. However, the presence of a considerable fraction of false positives is a bottleneck in realizing this potential. There have been continuous efforts to utilize complementary resources for scoring confidence of PPIs in a manner that false positive interactions get a low confidence score. Gene Ontology (GO), a taxonomy of biological terms to represent the properties of gene products and their relations, has been widely used for this purpose. We utilize GO to introduce a new set of specificity measures: Relative Depth Specificity (RDS), Relative Node-based Specificity (RNS), and Relative Edge-based Specificity (RES), leading to a new family of similarity measures. We use these similarity measures to obtain a confidence score for each PPI. We evaluate the new measures using four different benchmarks. We show that all the three measures are quite effective. Notably, RNS and RES more effectively distinguish true PPIs from false positives than the existing alternatives. RES also shows a robust set-discriminating power and can be useful for protein functional clustering as well.

A Collection of Benchmark Data Sets for Knowledge Graph-based Similarity in the Biomedical Domain

“…This tool enables the comparison of new GO-based semantic similarity measures against previously published ones considering their relation to sequence, Pfam ( 34 ) and Enzyme Commission (EC) ( 35 ) number similarity. CESSM was released in 2009 and updated in 2014, and since then it has been widely used by the community, being adopted to evaluate over 25 novel semantic similarity measures developed through different methods, with more recent ones focusing on common information content (IC)-based metrics ( 36 ) but also based on vector representations/graph embeddings ( 37 ). CESSM was built as a web-based tool to support the automatic comparison against the benchmark data.…”

Impact of the Continuous Evolution of Gene Ontology on the Performance of Similarity Measures for Scoring Confidence of Protein Interactions

“…Resnik and TCSS with MAX strategy have been shown to be the best SSMs for scoring confidence of PPIs by several studies [12,16,31,50]. We further consider three recently proposed SSMs (RDS, RNS, and RES) by Paul and Anand [28]. It has been shown by the authors that the new SSMs, particularly RES, outperform Resnik, and TCSS in many cases for scoring confidence of PPIs.…”

“…Nine different Bioconductor versions (3.0, 3.1, 3.2, 3.3, 3.4, 3.5, 3.6, 3.7, and 3.8) of the aforementioned three packages are considered. For RDS, RNS, and RES, we use the R script provided by the authors [28]. We maintain versions of all R packages so that they use the same GO and corresponding annotations.…”

“…But GO is an incomplete and continuously evolving ontology. Recently, a new family of SSMs (RDS, RNS, and RES) has been proposed by considering an alternative definition of specificity that is not based on the whole ontology, instead of exploiting the properties of subgraphs containing only ancestors and descendants of a given term to minimize the impact of the continuous evolution [28].…”

“…The same set of 11 yeast (S. cerevisiae) KEGG pathways (Table 1), as in [28], is selected in the evaluation. While selecting pathways, the authors consider the following two issues: (1) the functionality of the pathways should be as diverse as possible, and (2) the number of genes should not be too low or too high so that a trade-off between functional diversity and computational time can be maintained.…”