Propagation Network Algorithm Research Articles

A feature extraction framework for discovering pan‐cancer driver genes based on multi‐omics data

AbstractThe identification of tumor driver genes facilitates accurate cancer diagnosis and treatment, playing a key role in precision oncology, along with gene signaling, regulation, and their interaction with protein complexes. To tackle the challenge of distinguishing driver genes from a large number of genomic data, we construct a feature extraction framework for discovering pan‐cancer driver genes based on multi‐omics data (mutations, gene expression, copy number variants, and DNA methylation) combined with protein–protein interaction (PPI) networks. Using a network propagation algorithm, we mine functional information among nodes in the PPI network, focusing on genes with weak node information to represent specific cancer information. From these functional features, we extract distribution features of pan‐cancer data, pan‐cancer TOPSIS features of functional features using the ideal solution method, and SetExpan features of pan‐cancer data from the gene functional features, a method to rank pan‐cancer data based on the average inverse rank. These features represent the common message of pan‐cancer. Finally, we use the lightGBM classification algorithm for gene prediction. Experimental results show that our method outperforms existing methods in terms of the area under the check precision‐recall curve (AUPRC) and demonstrates better performance across different PPI networks. This indicates our framework’s effectiveness in predicting potential cancer genes, offering valuable insights for the diagnosis and treatment of tumors.

Network propagation unveils molecular mechanisms modulated in Alzheimer’s Disease patients treated by the complex plasma fraction GRF6019

AbstractBackgroundOur pre‐clinical work provided evidence of the interest of plasma fraction treatment to reverse age‐related cognitive decline. One of the most potent plasma fractions, GRF6019 was tested on patients with Alzheimer’s disease (AD). GRF6019 was safe and well tolerated, and patients experienced no cognitive decline and minimal functional decline. However, understanding these responses requires large‐scale mapping of the underlying molecular mechanisms. Recent affinity‐based approaches can measure levels of thousands of proteins in plasma but the increased signal‐to‐noise ratio and the large‐scale analysis of the proteomics changes require sophisticated bioinformatic techniques. To this end, we applied network propagation methods and combined protein‐protein interaction (PPI) networks, known AD‐associated proteins and GRF6019 induced protein changes to shed the light on the molecular mechanisms underlying patients’ responses to GRF6019 treatment.MethodPhase 2 clinical trials with repeated infusions of the plasma fraction GRF6019 were conducted in patients with AD (GRF6019‐201 n = 40 and GRF6019‐202 n = 26). Proteomic measurements from 136 plasma samples across three timepoints were performed by SomaLogic and Olink platforms. GRF6019‐induced proteomics changes were expressed by a Modulation Score (MS) for each protein. This MS captures the amplitude of the proteomics changes and the consistency of these changes within and between trials. Network propagation modeled the spread of MS on the PPI network. The network propagation algorithm we applied was the fusion of Hierarchical HotNet (permutation part) and NetCore (coreness normalization part). The algorithm highlighted the perturbed proteins which are close in the PPI network topology to proteins with high MS.ResultHighly perturbed proteins were more strongly associated with AD and other age‐related diseases (e.g.: AMD, Ischemic stroke) than proteins with high MS. Several pathways involved in AD were found to be associated to the perturbed proteins, such as complement and monocyte activation, or regulation of neurogenesis. Furthermore, the high overlap between proteins perturbed by GRF6019 and known AD‐associated proteins indicates the importance of indirect links between the GRF6019 treatment and AD.ConclusionNetwork propagation procedure highlights the importance of interactome‐based analysis in understanding complex plasma proteomics signals. These results suggest that the treatment of AD patients with a complex plasma fraction modulates AD‐related biological pathways.

Network embedding framework for driver gene discovery by combining functional and structural information

Comprehensive analysis of multiple data sets can identify potential driver genes for various cancers. In recent years, driver gene discovery based on massive mutation data and gene interaction networks has attracted increasing attention, but there is still a need to explore combining functional and structural information of genes in protein interaction networks to identify driver genes. Therefore, we propose a network embedding framework combining functional and structural information to identify driver genes. Firstly, we combine the mutation data and gene interaction networks to construct mutation integration network using network propagation algorithm. Secondly, the struc2vec model is used for extracting gene features from the mutation integration network, which contains both gene's functional and structural information. Finally, machine learning algorithms are utilized to identify the driver genes. Compared with the previous four excellent methods, our method can find gene pairs that are distant from each other through structural similarities and has better performance in identifying driver genes for 12 cancers in the cancer genome atlas. At the same time, we also conduct a comparative analysis of three gene interaction networks, three gene standard sets, and five machine learning algorithms. Our framework provides a new perspective for feature selection to identify novel driver genes.

DTSEA: A network-based drug target set enrichment analysis method for drug repurposing against COVID-19

The Coronavirus Disease 2019 (COVID-19) pandemic is still wreaking havoc worldwide. Therefore, the urgent need for efficient treatments pushes researchers and clinicians into screening effective drugs. Drug repurposing may be a promising and time-saving strategy to identify potential drugs against this disease. Here, we developed a novel computational approach, named Drug Target Set Enrichment Analysis (DTSEA), to identify potent drugs against COVID-19. DTSEA first mapped the disease-related genes into a gene functional interaction network, and then it used a network propagation algorithm to rank all genes in the network by calculating the network proximity of genes to disease-related genes. Finally, an enrichment analysis was performed on drug target sets to prioritize disease-candidate drugs. It was shown that the top three drugs predicted by DTSEA, including Ataluren, Carfilzomib, and Aripiprazole, were significantly enriched in the immune response pathways indicating the potential for use as promising COVID-19 inhibitors. In addition to these drugs, DTSEA also identified several drugs (such as Remdesivir and Olumiant), which have obtained emergency use authorization (EUA) for COVID-19. These results indicated that DTSEA could effectively identify the candidate drugs for COVID-19, which will help to accelerate the development of drugs for COVID-19. We then performed several validations to ensure the reliability and validity of DTSEA, including topological analysis, robustness analysis, and prediction consistency. Collectively, DTSEA successfully predicted candidate drugs against COVID-19 with high accuracy and reliability, thus making it a formidable tool to identify potential drugs for a specific disease and facilitate further investigation.

IATMEcell: identification of abnormal tumor microenvironment cells to predict the clinical outcomes in cancer based on cell-cell crosstalk network.

Interactions between Tumor microenvironment (TME) cells shape the unique growth environment, sustaining tumor growth and causing the immune escape of tumor cells. Nonetheless, no studies have reported a systematic analysis of cellular interactions in the identification of cancer-related TME cells. Here, we proposed a novel network-based computational method, named as iATMEcell, to identify the abnormal TME cells associated with the biological outcome of interest based on a cell-cell crosstalk network. In the method, iATMEcell first manually collected TME cell types from multiple published studies and obtained their corresponding gene signatures. Then, a weighted cell-cell crosstalk network was constructed in the context of a specific cancer bulk tissue transcriptome data, where the weight between cells reflects both their biological function similarity and the transcriptional dysregulated activities of gene signatures shared by them. Finally, it used a network propagation algorithm to identify significantly dysregulated TME cells. Using the cancer genome atlas (TCGA) Bladder Urothelial Carcinoma training set and two independent validation sets, we illustrated that iATMEcell could identify significant abnormal cells associated with patient survival and immunotherapy response. iATMEcell was further applied to a pan-cancer analysis, which revealed that four common abnormal immune cells play important roles in the patient prognosis across multiple cancer types. Collectively, we demonstrated that iATMEcell could identify potentially abnormal TME cells based on a cell-cell crosstalk network, which provided a new insight into understanding the effect of TME cells in cancer. iATMEcell is developed as an R package, which is freely available on GitHub (https://github.com/hanjunwei-lab/iATMEcell).

NetMix2: A Principled Network Propagation Algorithm for Identifying Altered Subnetworks.

A standard paradigm in computational biology is to leverage interaction networks as prior knowledge in analyzing high-throughput biological data, where the data give a score for each vertex in the network. One classical approach is the identification of altered subnetworks, or subnetworks of the interaction network that have both outlier vertex scores and a defined network topology. One class of algorithms for identifying altered subnetworks search for high-scoring subnetworks in subnetwork families with simple topological constraints, such as connected subnetworks, and have sound statistical guarantees. A second class of algorithms employ network propagation-the smoothing of vertex scores over the network using a random walk or diffusion process-and utilize the global structure of the network. However, network propagation algorithms often rely on ad hoc heuristics that lack a rigorous statistical foundation. In this work, we unify the subnetwork family and network propagation approaches by deriving the propagation family, a subnetwork family that approximates the sets of vertices ranked highly by network propagation approaches. We introduce NetMix2, a principled algorithm for identifying altered subnetworks from a wide range of subnetwork families. When using the propagation family, NetMix2 combines the advantages of the subnetwork family and network propagation approaches. NetMix2 outperforms other methods, including network propagation on simulated data, pan-cancer somatic mutation data, and genome-wide association data from multiple human diseases.

PhenoGeneRanker: Gene and Phenotype Prioritization Using Multiplex Heterogeneous Networks.

Uncovering genotype-phenotype relationships is a fundamental challenge in genomics. Gene prioritization is an important step for this endeavor to make a short manageable list from a list of thousands of genes coming from high-throughput studies. Network propagation methods are promising and state of the art methods for gene prioritization based on the premise that functionally related genes tend to be close to each other in the biological networks. Recently, we introduced PhenoGeneRanker, a network-propagation algorithm for multiplex heterogeneous networks. PhenoGeneRanker allows multi-layer gene and phenotype networks. It also calculates empirical p values of gene and phenotype ranks using random stratified sampling of seeds of genes and phenotypes based on their connectivity degree in the network. In this study, we introduce the PhenoGeneRanker Bioconductor package and its application to multi-omics rat genome datasets to rank hypertension disease-related genes and strains. We showed that PhenoGeneRanker performed better to rank hypertension disease-related genes using multiplex gene networks than aggregated gene networks. We also showed that PhenoGeneRanker performed better to rank hypertension disease-related strains using multiplex phenotype network than single or aggregated phenotype networks. We performed a rigorous hyperparameter analysis and, finally showed that Gene Ontology (GO) enrichment of statistically significant top-ranked genes resulted in hypertension disease-related GO terms.

Network-Based Approaches for Disease-Gene Association Prediction Using Protein-Protein Interaction Networks

Genome-wide association studies (GWAS) can be used to infer genome intervals that are involved in genetic diseases. However, investigating a large number of putative mutations for GWAS is resource- and time-intensive. Network-based computational approaches are being used for efficient disease-gene association prediction. Network-based methods are based on the underlying assumption that the genes causing the same diseases are located close to each other in a molecular network, such as a protein-protein interaction (PPI) network. In this survey, we provide an overview of network-based disease-gene association prediction methods based on three categories: graph-theoretic algorithms, machine learning algorithms, and an integration of these two. We experimented with six selected methods to compare their prediction performance using a heterogeneous network constructed by combining a genome-wide weighted PPI network, an ontology-based disease network, and disease-gene associations. The experiment was conducted in two different settings according to the presence and absence of known disease-associated genes. The results revealed that HerGePred, an integrative method, outperformed in the presence of known disease-associated genes, whereas PRINCE, which adopted a network propagation algorithm, was the most competitive in the absence of known disease-associated genes. Overall, the results demonstrated that the integrative methods performed better than the methods using graph-theory only, and the methods using a heterogeneous network performed better than those using a homogeneous PPI network only.

Optimizing network propagation for multi-omics data integration.

Network propagation refers to a class of algorithms that integrate information from input data across connected nodes in a given network. These algorithms have wide applications in systems biology, protein function prediction, inferring condition-specifically altered sub-networks, and prioritizing disease genes. Despite the popularity of network propagation, there is a lack of comparative analyses of different algorithms on real data and little guidance on how to select and parameterize the various algorithms. Here, we address this problem by analyzing different combinations of network normalization and propagation methods and by demonstrating schemes for the identification of optimal parameter settings on real proteome and transcriptome data. Our work highlights the risk of a 'topology bias' caused by the incorrect use of network normalization approaches. Capitalizing on the fact that network propagation is a regularization approach, we show that minimizing the bias-variance tradeoff can be utilized for selecting optimal parameters. The application to real multi-omics data demonstrated that optimal parameters could also be obtained by either maximizing the agreement between different omics layers (e.g. proteome and transcriptome) or by maximizing the consistency between biological replicates. Furthermore, we exemplified the utility and robustness of network propagation on multi-omics datasets for identifying ageing-associated genes in brain and liver tissues of rats and for elucidating molecular mechanisms underlying prostate cancer progression. Overall, this work compares different network propagation approaches and it presents strategies for how to use network propagation algorithms to optimally address a specific research question at hand.

Predicting anticancer hyperfoods with graph convolutional networks

BackgroundRecent efforts in the field of nutritional science have allowed the discovery of disease-beating molecules within foods based on the commonality of bioactive food molecules to FDA-approved drugs. The pioneering work in this field used an unsupervised network propagation algorithm to learn the systemic-wide effect on the human interactome of 1962 FDA-approved drugs and a supervised algorithm to predict anticancer therapeutics using the learned representations. Then, a set of bioactive molecules within foods was fed into the model, which predicted molecules with cancer-beating potential.The employed methodology consisted of disjoint unsupervised feature generation and classification tasks, which can result in sub-optimal learned drug representations with respect to the classification task. Additionally, due to the disjoint nature of the tasks, the employed approach proved cumbersome to optimize, requiring testing of thousands of hyperparameter combinations and significant computational resources.To overcome the technical limitations highlighted above, we represent each drug as a graph (human interactome) with its targets as binary node features on the graph and formulate the problem as a graph classification task. To solve this task, inspired by the success of graph neural networks in graph classification problems, we use an end-to-end graph neural network model operating directly on the graphs, which learns drug representations to optimize model performance in the prediction of anticancer therapeutics.ResultsThe proposed model outperforms the baseline approach in the anticancer therapeutic prediction task, achieving an F1 score of 67.99%±2.52% and an AUPR of 73.91%±3.49%. It is also shown that the model is able to capture knowledge of biological pathways to predict anticancer molecules based on the molecules’ effects on cancer-related pathways.ConclusionsWe introduce an end-to-end graph convolutional model to predict cancer-beating molecules within food. The introduced model outperforms the existing baseline approach, and shows interpretability, paving the way to the future of a personalized nutritional science approach allowing the development of nutrition strategies for cancer prevention and/or therapeutics.

Abstract 4891: Hepatitis B virus remodels host protein interaction networks to generate distinct cellular dependencies

Abstract Hepatocellular carcinoma (HCC) is the second leading cause of cancer death worldwide. Advanced HCC has proven particularly difficult to treat because of a scarcity of clear genetic drivers of cancer progression; thus, there are currently no predictive markers that guide HCC therapy. HCC arises in the context of co-morbid hepatitis due to hepatitis B virus (HBV), hepatitis C (HCV) or fatty liver disease. We hypothesize that protein-protein interactions (PPIs) between viral proteins and HCC genes may contribute to tumor initiation and maintenance. In order to characterize these PPIs, we performed affinity purification - mass spectrometry (APMS), defining 145 HBV/host PPIs including known and novel interacting partners. We next used a network propagation algorithm to identify host genes and protein complexes that were preferentially mutated in the absence of HBV infection. HBV is a small DNA virus, with 4 genes of which only one has enzymatic activity, raising a question as to how HBV interaction modifies host behavior. Using AP-MS of host proteins, we found that the HBV X protein (HBx) remodels multiple host protein complexes through direct interaction. These physical effects on complex components result in distinct biochemical behavior from the CRL4 E3 ubiquitin ligase complex as well as the phosphatase PP2A, as determined through global phosphoproteomics and ubiquitin analysis. We show that this remodeling driven by HBx substantially changes cellular protein turnover and downstream signaling dynamics. We followed this up with assessments of cellular viability and proliferation in response to pharmacological inhibition or CRISPRi-based knockdown of HBx effectors. Our data support a model where HBV proteins alter the components and behavior of key regulatory protein complexes in the cell, altering tumor behavior and raising the possibility of precision therapeutics for HCC. Citation Format: John D. Gordan, Adriana Pitea, Manon Eckhardt, Gwendolyn Jang, Rigney E. Turnham, Alex L M. Choi, John Von Dollen, Huat C. Lim, Elizabeth F. Thayer, R. Katie Kelley, Danielle L. Swaney, Wei Zhang, Fabian J. Theis, Trey Ideker, Nevan J. Krogan. Hepatitis B virus remodels host protein interaction networks to generate distinct cellular dependencies [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 4891.

Multi-scale top-down approach for modelling epileptic protein-protein interaction network analysis to identify driver nodes and pathways

Protein - Protein Interaction Network (PPIN) analysis unveils molecular level mechanisms involved in disease condition. To explore the complex regulatory mechanisms behind epilepsy and to address the clinical and biological issues of epilepsy, in silico techniques are feasible in a cost- effective manner. In this work, a hierarchical procedure to identify influential genes and regulatory pathways in epilepsy prognosis is proposed. To obtain key genes and pathways causing epilepsy, integration of two benchmarked datasets which are exclusively devoted for complex disorders is done as an initial step. Using STRING database, PPIN is constructed for modelling protein-protein interactions. Further, key interactions are obtained from the established PPIN using network centrality measures followed by network propagation algorithm -Random Walk with Restart (RWR). The outcome of the method reveals some influential genes behind epilepsy prognosis, along with their associated pathways like PI3 kinase, VEGF signaling, Ras, Wnt signaling etc. In comparison with similar works, our results have shown improvement in identifying unique molecular functions, biological processes, gene co-occurrences etc. Also, CORUM provides an annotation for approximately 60% of similarity in human protein complexes with the obtained result. We believe that the formulated strategy can put-up the vast consideration of indigenous drugs towards meticulous identification of genes encoded by protein against several combinatorial disorders.

Prioritization of candidate cancer drugs based on a drug functional similarity network constructed by integrating pathway activities and drug activities.

Due to the speed, efficiency, relative risk, and lower costs compared to traditional drug discovery, the prioritization of candidate drugs for repurposing against cancers of interest has attracted the attention of experts in recent years. Herein, we present a powerful computational approach, termed prioritization of candidate drugs (PriorCD), for the prioritization of candidate cancer drugs based on a global network propagation algorithm and a drug–drug functional similarity network constructed by integrating pathway activity profiles and drug activity profiles. This provides a new approach to drug repurposing by first considering the drug functional similarities at the pathway level. The performance of PriorCD in drug repurposing was evaluated by using drug datasets of breast cancer and ovarian cancer. Cross‐validation tests on the drugs approved for the treatment of these cancers indicated that our approach can achieve area under receiver‐operating characteristic curve (AUROC) values greater than 0.82. Furthermore, literature searches validated our results, and comparison with other classical gene‐based repurposing methods indicated that our pathway‐level PriorCD is comparatively more effective at prioritizing candidate drugs with similar therapeutic effects. We hope that our study will be of benefit to the field of drug discovery. In order to expand the usage of PriorCD, a freely available R‐based package, PriorCD, has been developed to prioritize candidate anticancer drugs for drug repurposing.

Exploration of Target Synergy in Cancer Treatment by Cell-Based Screening Assay and Network Propagation Analysis.

Computational approaches have previously been introduced to predict compounds with activity against multiple targets or compound combinations with synergistic functional effects. By contrast, there are no computational studies available that explore combinations of targets that might act synergistically upon small molecule treatment. Herein, we introduce an approach designed to identify synergistic target pairs on the basis of cell-based screening data and compounds with known target annotations. The targets involved in forming synergistic pairs were analyzed through a novel network propagation algorithm for rationalizing possible common synergy mechanisms. This algorithm enabled further analysis of each synergistic target pair and the identification of "interactors", i.e., proteins with higher propagation scores than would be expected by adding the individual contributions of each target in the synergistic pair. We detected 137 synergistic target pairs including 51 unique targets. A global network analysis of these 51 targets made it possible to derive a subnetwork of proteins with significant synergy. Furthermore, interactors were identified for 87 synergistic target pairs upon individual analysis of the network propagation of each pair. These interactors were associated with pathways related to cancer and apoptosis, membrane transport, and steroid metabolism and provided possible explanations of synergistic effects.

Integrating Biological Networks for Drug Target Prediction and Prioritization.

Computational prediction of the clinical success or failure of a potential drug target for therapeutic use is a challenging problem. Novel network propagation algorithms that integrate heterogeneous biological networks are proving useful for drug target identification and prioritization. These approaches typically utilize a network describing relationships between targets, a method to disseminate the relevant information through the network, and a method to elucidate new associations between targets and diseases. Here, we utilize one such network propagation-based approach, DTINet, which starts with diffusion component analysis of networks of both potential drug targets and diseases. Then an inductive matrix completion algorithm is applied to identify novel disease targets based on their network topological similarities with known disease targets with successfully launched drugs. DTINet performed well as assessed with area under the precision-recall curve (AUPR=0.88±0.007) and area under the receiver operating characteristic curve (AUROC=0.86±0.008). These metrics improved when we combined data from multiple networks in the target space but reduced significantly when we used a more conservative method to define negative controls (AUPR=0.56±0.007, AUROC=0.57±0.007). We are optimistic that integration of more relevant and cleaner datasets and networks, careful calibration of model parameters, as well as algorithmic improvements will improve prediction accuracy. However, we also recognize that predicting drug targets that are likely to be successful is an extremely challenging problem due to its complex nature and sparsity of known disease targets.

Predicting disease-genes based on network information loss and protein complexes in heterogeneous network

Research on disease-related genes has always been a visible focus in the biological field. This study can help us to reveal the hidden mechanism of human diseases. Although many methods have been developed, the accuracy of identifying disease-genes remains challenging. We propose a computation-based method for predicting disease-genes. The investigations are inspired by the information loss model for evaluating similarities between network nodes and human protein complexes that reflect interactions between genes. Furthermore, we also combine other data such as lncRNA to construct a triple heterogeneous network and design a network propagation algorithm applied to the heterogeneous network (InLPCH). This algorithm effectively reduces the number of false positives in the biological networks when predicting disease-genes and combines the multiple propagation paths of the heterogeneous network to improve prediction accuracy. We conduct extensive experiments over disease-genes dataset. The InLPCH demonstrates high performance in comparison with six other state-of-the-art algorithms.

Genome-wide analysis of long intergenic non-coding RNAs in chickpea and their potential role in flower development.

Non-coding RNAs constitute a major portion of the transcriptome in most of eukaryotes. Long non-coding transcripts originating from the DNA segment present between the protein coding genes are termed as long intergenic non-coding RNAs (lincRNAs). Several evidences suggest the role of lincRNAs in regulation of various biological processes. In this study, we identified a total of 2248 lincRNAs in chickpea using RNA-seq data from eight successive stages of flower development and three vegetative tissues via an optimized pipeline. Different characteristic features of lincRNAs were studied and compared with those of predicted mRNAs in chickpea. Further, we utilized a method using network propagation algorithm to reveal the putative function of lincRNAs in plants. In total, at least 79% of the identified chickpea lincRNAs were assigned with a putative function. A comprehensive expression profiling revealed differential expression patterns and tissue specificity of lincRNAs in different stages of flower development in chickpea. In addition, potential lincRNAs-miRNA interactions were explored for the predicted lincRNAs in chickpea. These findings will pave the way for understanding the role of lincRNAs in the regulatory mechanism underlying flower development in chickpea and other legumes.

Network-Based Integration of Disparate Omic Data To Identify "Silent Players" in Cancer.

Development of high-throughput monitoring technologies enables interrogation of cancer samples at various levels of cellular activity. Capitalizing on these developments, various public efforts such as The Cancer Genome Atlas (TCGA) generate disparate omic data for large patient cohorts. As demonstrated by recent studies, these heterogeneous data sources provide the opportunity to gain insights into the molecular changes that drive cancer pathogenesis and progression. However, these insights are limited by the vast search space and as a result low statistical power to make new discoveries. In this paper, we propose methods for integrating disparate omic data using molecular interaction networks, with a view to gaining mechanistic insights into the relationship between molecular changes at different levels of cellular activity. Namely, we hypothesize that genes that play a role in cancer development and progression may be implicated by neither frequent mutation nor differential expression, and that network-based integration of mutation and differential expression data can reveal these “silent players”. For this purpose, we utilize network-propagation algorithms to simulate the information flow in the cell at a sample-specific resolution. We then use the propagated mutation and expression signals to identify genes that are not necessarily mutated or differentially expressed genes, but have an essential role in tumor development and patient outcome. We test the proposed method on breast cancer and glioblastoma multiforme data obtained from TCGA. Our results show that the proposed method can identify important proteins that are not readily revealed by molecular data, providing insights beyond what can be gleaned by analyzing different types of molecular data in isolation.

A novel method for identifying disease associated protein complexes based on functional similarity protein complex networks.

BackgroundProtein complexes formed by non-covalent interaction among proteins play important roles in cellular functions. Computational and purification methods have been used to identify many protein complexes and their cellular functions. However, their roles in terms of causing disease have not been well discovered yet. There exist only a few studies for the identification of disease-associated protein complexes. However, they mostly utilize complicated heterogeneous networks which are constructed based on an out-of-date database of phenotype similarity network collected from literature. In addition, they only apply for diseases for which tissue-specific data exist.MethodsIn this study, we propose a method to identify novel disease-protein complex associations. First, we introduce a framework to construct functional similarity protein complex networks where two protein complexes are functionally connected by either shared protein elements, shared annotating GO terms or based on protein interactions between elements in each protein complex. Second, we propose a simple but effective neighborhood-based algorithm, which yields a local similarity measure, to rank disease candidate protein complexes.ResultsComparing the predictive performance of our proposed algorithm with that of two state-of-the-art network propagation algorithms including one we used in our previous study, we found that it performed statistically significantly better than that of these two algorithms for all the constructed functional similarity protein complex networks. In addition, it ran about 32 times faster than these two algorithms. Moreover, our proposed method always achieved high performance in terms of AUC values irrespective of the ways to construct the functional similarity protein complex networks and the used algorithms. The performance of our method was also higher than that reported in some existing methods which were based on complicated heterogeneous networks. Finally, we also tested our method with prostate cancer and selected the top 100 highly ranked candidate protein complexes. Interestingly, 69 of them were evidenced since at least one of their protein elements are known to be associated with prostate cancer.ConclusionsOur proposed method, including the framework to construct functional similarity protein complex networks and the neighborhood-based algorithm on these networks, could be used for identification of novel disease-protein complex associations.Electronic supplementary materialThe online version of this article (doi:10.1186/s13015-015-0044-6) contains supplementary material, which is available to authorized users.

Robust and efficient identification of biomarkers by classifying features on graphs

A central problem in biomarker discovery from large-scale gene expression or single nucleotide polymorphism (SNP) data is the computational challenge of taking into account the dependence among all the features. Methods that ignore the dependence usually identify non-reproducible biomarkers across independent datasets. We introduce a new graph-based semi-supervised feature classification algorithm to identify discriminative disease markers by learning on bipartite graphs. Our algorithm directly classifies the feature nodes in a bipartite graph as positive, negative or neutral with network propagation to capture the dependence among both samples and features (clinical and genetic variables) by exploring bi-cluster structures in a graph. Two features of our algorithm are: (1) our algorithm can find a global optimal labeling to capture the dependence among all the features and thus, generates highly reproducible results across independent microarray or other high-thoughput datasets, (2) our algorithm is capable of handling hundreds of thousands of features and thus, is particularly useful for biomarker identification from high-throughput gene expression and SNP data. In addition, although designed for classifying features, our algorithm can also simultaneously classify test samples for disease prognosis/diagnosis. We applied the network propagation algorithm to study three large-scale breast cancer datasets. Our algorithm achieved competitive classification performance compared with SVMs and other baseline methods, and identified several markers with clinical or biological relevance with the disease. More importantly, our algorithm also identified highly reproducible marker genes and enriched functions from the independent datasets. Supplementary results and source code are available at http://compbio.cs.umn.edu/Feature_Class. Supplementary data are available at Bioinformatics online.