An Integrative -omics Approach to Identify Functional Sub-Networks in Human Colorectal Cancer

University of Illinois at Urbana-Champaign, United States of America
PLoS Computational Biology (Impact Factor: 4.62). 01/2010; 6(1):e1000639. DOI: 10.1371/journal.pcbi.1000639
Source: PubMed


Emerging evidence indicates that gene products implicated in human cancers often cluster together in "hot spots" in protein-protein interaction (PPI) networks. Additionally, small sub-networks within PPI networks that demonstrate synergistic differential expression with respect to tumorigenic phenotypes were recently shown to be more accurate classifiers of disease progression when compared to single targets identified by traditional approaches. However, many of these studies rely exclusively on mRNA expression data, a useful but limited measure of cellular activity. Proteomic profiling experiments provide information at the post-translational level, yet they generally screen only a limited fraction of the proteome. Here, we demonstrate that integration of these complementary data sources with a "proteomics-first" approach can enhance the discovery of candidate sub-networks in cancer that are well-suited for mechanistic validation in disease. We propose that small changes in the mRNA expression of multiple genes in the neighborhood of a protein-hub can be synergistically associated with significant changes in the activity of that protein and its network neighbors. Further, we hypothesize that proteomic targets with significant fold change between phenotype and control may be used to "seed" a search for small PPI sub-networks that are functionally associated with these targets. To test this hypothesis, we select proteomic targets having significant expression changes in human colorectal cancer (CRC) from two independent 2-D gel-based screens. Then, we use random walk based models of network crosstalk and develop novel reference models to identify sub-networks that are statistically significant in terms of their functional association with these proteomic targets. Subsequently, using an information-theoretic measure, we evaluate synergistic changes in the activity of identified sub-networks based on genome-wide screens of mRNA expression in CRC. Cross-classification experiments to predict disease class show excellent performance using only a few sub-networks, underwriting the strength of the proposed approach in discovering relevant and reproducible sub-networks.

Full-text preview

Available from: PubMed Central
  • Source
    • "Weinberg et al. summarized the first and next generation of cancer hallmarks to expand the current understanding of the basic mechanisms of cancer [2,3]. Recently, due to the scale up in high throughput data, availability of integrated OMICS data, and various advanced statistical analysis methods, many novel systems biology approaches have been employed to reveal the deeper underlying systematic mechanisms of various cancers456. Traditional computer-aided drug design (CADD) focuses on a single target for therapy, such as Src, FAK, and EGFR in the case of cancer [7,8]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Computer-aided drug design has a long history of being applied to discover new molecules to treat various cancers, but it has always been focused on single targets. The development of systems biology has let scientists reveal more hidden mechanisms of cancers, but attempts to apply systems biology to cancer therapies remain at preliminary stages. Our lab has successfully developed various systems biology models for several cancers. Based on these achievements, we present the first attempt to combine multiple-target therapy with systems biology. In our previous study, we identified 28 significant proteins--i.e., common core network markers--of four types of cancers as house-keeping proteins of these cancers. In this study, we ranked these proteins by summing their carcinogenesis relevance values (CRVs) across the four cancers, and then performed docking and pharmacophore modeling to do virtual screening on the NCI database for anti-cancer drugs. We also performed pathway analysis on these proteins using Panther and MetaCore to reveal more mechanisms of these cancer house-keeping proteins. We designed several approaches to discover targets for multiple-target cocktail therapies. In the first one, we identified the top 20 drugs for each of the 28 cancer house-keeping proteins, and analyzed the docking pose to further understand the interaction mechanisms of these drugs. After screening for duplicates, we found that 13 of these drugs could target 11 proteins simultaneously. In the second approach, we chose the top 5 proteins with the highest summed CRVs and used them as the drug targets. We built a pharmacophore and applied it to do virtual screening against the Life-Chemical library for anti-cancer drugs. Based on these results, wet-lab bio-scientists could freely investigate combinations of these drugs for multiple-target therapy for cancers, in contrast to the traditional single target therapy. Combination of systems biology with computer-aided drug design could help us develop novel drug cocktails with multiple targets. We believe this will enhance the efficiency of therapeutic practice and lead to new directions for cancer therapy.
    Full-text · Article · Dec 2015 · BMC Medical Genomics
  • Source
    • "It is believed that dynamic alternations of complex interaction networks and molecular sub-networks can represent and influence responses of cells or organs to real-time changed microenvironment [10-12]. Thus, identification and validation of interaction networks and network biomarkers, especially at the protein level, become critical to develop disease-specific biomarkers for monitoring disease occurrence, progression or treatment efficacy [13-15]. The present review headlights network biomarkers, interaction networks, dynamical network biomarkers, with special focus on respiratory diseases, with an emphasis to integrate bioinformatics-based screening of biomarkers, network biomarker, dynamic network biomarkers with clinical informatics and phenotypes and establish a systems biomedicine-evidenced disease-specific dynamic network biomarkers "
    [Show abstract] [Hide abstract]
    ABSTRACT: Identification and validation of interaction networks and network biomarkers have become more critical and important in the development of disease-specific biomarkers, which are functionally changed during disease development, progression or treatment. The present review headlined the definition, significance, research and potential application for network biomarkers, interaction networks and dynamical network biomarkers (DNB). Disease-specific interaction networks, network biomarkers, or DNB have great significance in the understanding of molecular pathogenesis, risk assessment, disease classification and monitoring, or evaluations of therapeutic responses and toxicities. Protein-based DNB will provide more information to define the differences between the normal and pre-disease stages, which might point to early diagnosis for patients. Clinical bioinformatics should be a key approach to the identification and validation of disease-specific biomarkers.
    Full-text · Article · Jun 2014 · Clinical and Translational Medicine
  • Source
    • "Protein network and mRNA profiles can be integrated to identify subnetwork biomarkers, that is, highly connected genes of a subnetwork whose sum of expression can be a marker of a disease state. There are several network-based approaches for identifying disease genes and protein interaction subnetworks which are disease signatures [86–88]. The application of a network analysis to metabolic PET (positron emission tomography) data obtained from patients with Parkinson's disease resulted in the identification and validation of two distinct spatial covariance patterns associated with the motor and cognitive manifestations of the disease [89]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: one is the increasing capabilities of the computers and software tools from terabytes to petabytes and beyond, and the other is the advancement in high-throughput molecular biology producing piles of data related to genomes, transcriptomes, proteomes, metabolomes, interactomes, and so on. Biology has become a data intensive science and as a consequence biology and computer science have become complementary to each other bridged by other branches of science such as statistics, mathematics, physics, and chemistry. The combination of versatile knowledge has caused the advent of big-data biology, network biology, and other new branches of biology. Network biology for instance facilitates the system-level understanding of the cell or cellular components and subprocesses. It is often also referred to as systems biology. The purpose of this field is to understand organisms or cells as a whole at various levels of functions and mechanisms. Systems biology is now facing the challenges of analyzing big molecular biological data and huge biological networks. This review gives an overview of the progress in big-data biology, and data handling and also introduces some applications of networks and multivariate analysis in systems biology.
    Full-text · Article · May 2014 · BioMed Research International
Show more