Abstract
Current multiomics assay platforms facilitate systematic identification of functional entities that are mappable in a biological network, and computational methods that are better able to detect densely connected clusters of signals within a biological network are considered increasingly important. One of the most famous algorithms for detecting network subclusters is Molecular Complex Detection (MCODE). MCODE, however, is limited in simultaneous analyses of multiple, large-scale data sets, since it runs on the Cytoscape platform, which requires extensive computational resources and has limited coding flexibility. In the present study, we implemented the MCODE algorithm in R programming language and developed a related package, which we called MCODER. We found the MCODER package to be particularly useful in analyzing multiple omics data sets simultaneously within the R framework. Thus, we applied MCODER to detect pharmacologically tractable protein-protein interactions selectively elevated in molecular subtypes of ovarian and colorectal tumors. In doing so, we found that a single molecular subtype representing epithelial-mesenchymal transition in both cancer types exhibited enhanced production of the collagen-integrin protein complex. These results suggest that tumors of this molecular subtype could be susceptible to pharmacological inhibition of integrin signaling.
Highlights
Biological functions often arise from multisubunit protein complexes, rather than a single, isolated protein [1, 2]
A number of algorithms have been developed over the past decade to tackle this problem, including the Markov Cluster Algorithm (MCL) [6], Molecular Complex Detection (MCODE) [7], DPClus [8], Affinity Propagation Clustering (APC) [9], Clustering based on Maximal Clique (CMC) [10], ClusterMaker [11], and Clustering with Overlapping Neighborhood Expansion (ClusterONE) [12]
We implemented the network clustering algorithm MCODE into the R software environment and demonstrated that the MCODER package saves computational resources and time, making it suited for analyzing multiple omics data sets
Summary
Biological functions often arise from multisubunit protein complexes, rather than a single, isolated protein [1, 2]. Because of biological and technical errors, these methods are hindered by a limited signal-to-noise ratio, rendering them vulnerable to high rates of false positives and false negatives; when discovered hits represent a single gene, protein, and so forth In this regard, codiscovery of hits for multiple subunits of a protein complex in an experimental condition helps mutually support the significance of such findings [4]. A number of algorithms have been developed over the past decade to tackle this problem, including the Markov Cluster Algorithm (MCL) [6], Molecular Complex Detection (MCODE) [7], DPClus [8], Affinity Propagation Clustering (APC) [9], Clustering based on Maximal Clique (CMC) [10], ClusterMaker [11], and Clustering with Overlapping Neighborhood Expansion (ClusterONE) [12] Many of these algorithms have been implemented in various Cytoscape applications (CytoCluster, ClusterViz [13], and ClusterMaker [11]), as well as in java-based applications (C-DEVA [14]). Cytoscape is a javabased, open source, bioinformatics software platform with
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
