A thorough analysis of the contribution of experimental, derived and sequence-based predicted protein-protein interactions for functional annotation of proteins

Abstract

<jats:p>Physical interaction between two proteins is strong evidence that the proteins are involved in the same biological process, making Protein-Protein Interaction (PPI) networks a valuable data resource for predicting the cellular functions of proteins. However, PPI networks are largely incomplete for non-model species. Here, we tested to what extened these incomplete networks are still useful for genome-wide function prediction. We used two network-based classifiers to predict Biological Process Gene Ontology terms from protein interaction data in four species: Saccharomyces cerevisiae, Escherichia coli, Arabidopsis thaliana and Solanum lycopersicum (tomato). The classifiers had reasonable performance in the well-studied yeast, but performed poorly in the other species. We showed that this poor performance can be considerably improved by adding edges predicted from various data sources, such as text mining, and that associations from the STRING database are more useful than interactions predicted by a neural network from sequence-based features.</jats:p>

Bio-JOIE: Joint Representation Learning of Biological Knowledge Bases

“…Specifically, gene ontology (GO) [10,16] is the most widely used resource for gene function annotation; STRING [29], PDB [2] and neXtProt [19] collect the knowledge accumulated from functional proteomic analysis; Expression Atlas [25] is a database facilitating the retrieval and analysis of gene expression studies. While those KBs provide the essential sources of knowledge for in silico research in the corresponding domains, such domain-specific knowledge is often sparse and costly to apprehend [21,30]. For example, PPI networks can be far from complete given the information supported by experimental results or suggested by computational inference [14,21].…”

“…While those KBs provide the essential sources of knowledge for in silico research in the corresponding domains, such domain-specific knowledge is often sparse and costly to apprehend [21,30]. For example, PPI networks can be far from complete given the information supported by experimental results or suggested by computational inference [14,21]. Makrodimitris et al [21] indicate that the numbers of protein-protein interactions in BIOGRID [24] for non-model organisms are far less than expected, specifically, there are only 107 interactions for tomato (Solanum lycopersicum) and 80 interactions for pig (Sus scrofa).…”

“…For example, PPI networks can be far from complete given the information supported by experimental results or suggested by computational inference [14,21]. Makrodimitris et al [21] indicate that the numbers of protein-protein interactions in BIOGRID [24] for non-model organisms are far less than expected, specifically, there are only 107 interactions for tomato (Solanum lycopersicum) and 80 interactions for pig (Sus scrofa). Evidently, relying on the KG from a single domain presents the risk of learning from limited and scarce information.…”