PPI network analyses of human WD40 protein family systematically reveal their tendency to assemble complexes and facilitate the complex predictions

Abstract

Background: WD40 repeat proteins constitute one of the largest families in eukaryotes, and widely participate in various fundamental cellular processes by interacting with other molecules. Based on individual WD40 proteins, previous work has demonstrated that their structural characteristics should confer great potential of interaction and complex formation, and has speculated that they may serve as hubs in the protein-protein interaction (PPI) network. However, what roles the whole family plays in organizing the PPI network, and whether this information can be utilized in complex prediction remain unclear. To address these issues, quantitative and systematic analyses of WD40 proteins from the perspective of PPI networks are highly required.

Results: In this work, we built two human PPI networks by using data sets with different confidence levels, and studied the network properties of the whole human WD40 protein family systematically. Our analyses have quantitatively confirmed that the human WD40 protein family, as a whole, tends to be hubs with an odds ratio of about 1.8 or greater, and the network decomposition has revealed that they are prone to enrich near the global center of the whole network with a fold change of two in the median k-values. By integrating expression profiles, we have further shown that WD40 hub proteins are inclined to be intramodular, which is indicative of complex assembling. Based on this information, we have further predicted 1674 potential WD40-associated complexes by choosing a clique-based method, which is more sensitive than others, and an indirect evaluation by co-expression scores has demonstrated its reliability.

Conclusions: At the systems level but not sporadic examples' level, this work has provided rich knowledge for better understanding WD40 proteins' roles in organizing the PPI network. These findings and predicted complexes can offer valuable clues for prioritizing candidates for further studies.

Fig. 1

Percentage of WD40 proteins in each k-core subnetwork during the decomposition of HC-PPI network. The percentages were obtained by dividing the number of WD40 proteins to that of total proteins in each k-core subnetwork

Fig. 2

Distributions of average PCCs of WD40 hubs, non-WD40 hubs, and randomized data in HC-PPI network. The solid lines in orange represent the WD40 hubs, the dotted lines in purple denote the non-WD40 hubs, and the longdash lines in blue represent the randomized data. The average PCCs are calculated by using both protein-level expression data (a) and RNA-level expression data (b)

Fig. 3

The number of reference complexes matched by the predicted complex sets at different ω scores. Different lines represent different predicted complex sets derived from different merging parameters. The ω at the X-axis denotes the score that determines whether a predicted complex matches a reference one. The Y-axis gives out the number of reference complexes matched by predicted ones at corresponding ω scores

Fig. 4

Distributions of the co-expression scores of predicted WD40-associated complexes, reference complexes, and decoy complexes. The orange solid line, the blue dotted line, and the black dashed line represent the distributions of co-expression scores of predicted WD40 complexes, reference complexes, and decoy complexes, respectively. The co-expression scores are calculated by using both the protein-level expression data (a) and the RNA-level expression data (b)

Fig. 5

Two examples of potential WD40 protein-associated complexes. The nodes connected by dark grey lines belong to predicted complexes, whereas the nodes connected by light grey lines represent the reference complexes. Nodes in light red are shared by the predicted complex and reference complex. a the predicted complex, core_209, superimposed with the reference complex CCT complex; b the predicted complex, core_5, superimposed with reference complex 19S proteasome