Selecting biologically informative genes in co-expression networks with a centrality score.

BACKGROUND: Measures of node centrality in biological networks are useful to detect genes with critical functional roles. In gene co-expression networks, highly connected genes (i.e., candidate hubs) have been associated with key disease-related pathways. Although different approaches to estimating gene centrality are available, their potential biological relevance in gene co-expression networks deserves further investigation. Moreover, standard measures of gene centrality focus on binary interaction networks, which may not always be suitable in the context of co-expression networks. Here, I also investigate a method that identifies potential biologically meaningful genes based on a weighted connectivity score and indicators of statistical relevance. RESULTS: The method enables a characterization of the strength and diversity of co-expression associations in the network. It outperformed standard centrality measures by highlighting more biologically informative genes in different gene co-expression networks and biological research domains. As part of the illustration of the gene selection potential of this approach, I present an application case in zebrafish heart regeneration. The proposed technique predicted genes that are significantly implicated in cellular processes required for tissue regeneration after injury. CONCLUSIONS: A method for selecting biologically informative genes from gene co-expression networks is provided, together with free open software.

Bioinformatics as a driver, not a passenger, of translational biomedical research: Perspectives from the 6th Benelux bioinformatics conference.

The 6th Benelux Bioinformatics Conference (BBC11) held in Luxembourg on 12 and 13 December 2011 attracted around 200 participants, including internationally-renowned guest speakers and more than 100 peer-reviewed submissions from 3 continents. Researchers from the public and private sectors convened at BBC11 to discuss advances and challenges in a wide spectrum of application areas. A key theme of the conference was the contribution of bioinformatics to enable and accelerate translational and clinical research. The BBC11 stressed the need for stronger collaborating efforts across disciplines and institutions. The demonstration of the clinical relevance of systems approaches and of next-generation sequencing-based measurement technologies are among the existing opportunities for increasing impact in translational research. Translational bioinformatics will benefit from research models that strike a balance between the importance of protecting intellectual property and the need to openly access scientific and technological advances. The full conference proceedings are freely available at http://www.bbc11.lu.

Information encoded in a network of inflammation proteins predicts clinical outcome after myocardial infarction.

BACKGROUND: Inflammation plays an important role in cardiac repair after myocardial infarction (MI). Nevertheless, the systems-level characterization of inflammation proteins in MI remains incomplete. There is a need to demonstrate the potential value of molecular network-based approaches to translational research. We investigated the interplay of inflammation proteins and assessed network-derived knowledge to support clinical decisions after MI. The main focus is the prediction of clinical outcome after MI. METHODS: We assembled My-Inflamome, a network of protein interactions related to inflammation and prognosis in MI. We established associations between network properties, disease biology and capacity to distinguish between prognostic categories. The latter was tested with classification models built on blood-derived microarray data from post-MI patients with different outcomes. This was followed by experimental verification of significant associations. RESULTS: My-Inflamome is organized into modules highly specialized in different biological processes relevant to heart repair. Highly connected proteins also tend to be high-traffic components. Such bottlenecks together with genes extracted from the modules provided the basis for novel prognostic models, which could not have been uncovered by standard analyses. Modules with significant involvement in transcriptional regulation are targeted by a small set of microRNAs. We suggest a new panel of gene expression biomarkers (TRAF2, SHKBP1 and UBC) with high discriminatory capability. Follow-up validations reported promising outcomes and motivate future research. CONCLUSION: This study enhances understanding of the interaction network that executes inflammatory responses in human MI. Network-encoded information can be translated into knowledge with potential prognostic application. Independent evaluations are required to further estimate the clinical relevance of the new prognostic genes.

Predictive integration of gene functional similarity and co-expression defines treatment response of endothelial progenitor cells.

BACKGROUND: Endothelial progenitor cells (EPCs) have been implicated in different processes crucial to vasculature repair, which may offer the basis for new therapeutic strategies in cardiovascular disease. Despite advances facilitated by functional genomics, there is a lack of systems-level understanding of treatment response mechanisms of EPCs. In this research we aimed to characterize the EPCs response to adenosine (Ado), a cardioprotective factor, based on the systems-level integration of gene expression data and prior functional knowledge. Specifically, we set out to identify novel biosignatures of Ado-treatment response in EPCs. RESULTS: The predictive integration of gene expression data and standardized functional similarity information enabled us to identify new treatment response biosignatures. Gene expression data originated from Ado-treated and -untreated EPCs samples, and functional similarity was estimated with Gene Ontology (GO)-based similarity information. These information sources enabled us to implement and evaluate an integrated prediction approach based on the concept of k-nearest neighbours learning (kNN). The method can be executed by expert- and data-driven input queries to guide the search for biologically meaningful biosignatures. The resulting integrated kNN system identified new candidate EPC biosignatures that can offer high classification performance (areas under the operating characteristic curve>0.8). We also showed that the proposed models can outperform those discovered by standard gene expression analysis. Furthermore, we report an initial independent in vitro experimental follow-up, which provides additional evidence of the potential validity of the top biosignature. CONCLUSION: Response to Ado treatment in EPCs can be accurately characterized with a new method based on the combination of gene co-expression data and GO-based similarity information. It also exploits the incorporation of human expert-driven queries as a strategy to guide the automated search for candidate biosignatures. The proposed biosignature improves the systems-level characterization of EPCs. The new integrative predictive modeling approach can also be applied to other phenotype characterization or biomarker discovery problems.

Drug-target network in myocardial infarction reveals multiple side effects of unrelated drugs.

The systems-level characterization of drug-target associations in myocardial infarction (MI) has not been reported to date. We report a computational approach that combines different sources of drug and protein interaction information to assemble the myocardial infarction drug-target interactome network (My-DTome). My-DTome comprises approved and other drugs interlinked in a single, highly-connected network with modular organization. We show that approved and other drugs may both be highly connected and represent network bottlenecks. This highlights influential roles for such drugs on seemingly unrelated targets and pathways via direct and indirect interactions. My-DTome modules are associated with relevant molecular processes and pathways. We find evidence that these modules may be regulated by microRNAs with potential therapeutic roles in MI. Different drugs can jointly impact a module. We provide systemic insights into cardiovascular effects of non-cardiovascular drugs. My-DTome provides the basis for an alternative approach to investigate new targets and multidrug treatment in MI.

Machine learning techniques for the automated classification of adhesin-like proteins in the human protozoan parasite Trypanosoma cruzi.

This paper reports on the evaluation of different machine learning techniques for the automated classification of coding gene sequences obtained from several organisms in terms of their functional role as adhesins. Diverse, biologically-meaningful, sequence-based features were extracted from the sequences and used as inputs to the in silico prediction models. Another contribution of this work is the generation of potentially novel and testable predictions about the surface protein DGF-1 family in Trypanosoma cruzi. Finally, these techniques are potentially useful for the automated annotation of known adhesin-like proteins from the trans-sialidase surface protein family in T. cruzi, the etiological agent of Chagas disease.

In silico, biologically-inspired modelling of genomic variation generation in surface proteins of Trypanosoma cruzi.

BACKGROUND: Protozoan parasites improve the likelihood of invading or adapting to the host through their capacity to present a large repertoire of surface molecules. The understanding of the mechanisms underlying the generation of antigenic diversity is crucial to aid in the development of therapies and the study of evolution. Despite advances driven by molecular biology and genomics, there is a need to gain a deeper understanding of key properties that may facilitate variation generation, models for explaining the role of genomic re-arrangements and the characterisation of surface protein families on the basis of their capacity to generate variation. Computer models may be implemented to explore, visualise and estimate the variation generation capacity of gene families in a dynamic fashion. In this paper we report the dynamic simulation of genomic variation using real T. cruzi coding sequences as inputs to a computational simulation system. The effects of random, multiple-point mutations and gene conversions on genomic variation generation were quantitatively estimated and visualised. Simulations were also implemented to investigate the potential role of pseudogenes as a source of antigenic variation in T. cruzi. RESULTS: Computational models of variation generation were applied to real coding sequences from surface proteins in T. cruzi: trans-sialidase-like proteins and putative surface protein dispersed gene family-1. In the simulations the sequences self-replicated, mutated and re-arranged during thousands of generations. Simulations were implemented for different mutation rates to estimate the relative robustness of the protein families in the face of DNA multiple-point mutations and sequence re-arrangements. The gene super-families and families showed distinguishing evolutionary responses, which may be used to characterise them on the basis of their capacity to generate variability. The simulations showed that sequences from T. cruzi nuclear genes tend to be relatively more robust against random, multiple-point mutations than those obtained from surface protein genes. Simulations also showed that a gene conversion model may act as an effective variation generation mechanism. Differential variation responses can be used to characterise the sequence groups under study. For example, unlike other families, sequences from the DGF1 family have the capacity to maximise variation at the amino acid level under relatively low mutation rates and through gene conversion. However, in relation to the other protein families, they exhibit more robust behaviour in response to more severe modifications through intra-family genomic sequence exchange. Independent simulations indicate that DGF1 pseudogenes might play a role in the generation of greater genomic variation in the DFG1 gene family through gene conversion under different experimental conditions. CONCLUSION: Digital, dynamic simulations may be implemented to characterise gene families on the basis of their capacity to generate variation in the face of genomic perturbations. Such simulations may be useful to explore antigenic variation mechanisms and hypotheses about robustness at the genomic level. This investigation illustrated how sequences derived from surface protein genes and computer simulations can be used to investigate variation generation mechanisms. Such in silico experiments of self-replicating sequences undergoing random mutations and genomic re-arrangements can offer insights into the diversity generation potential of the genes under study. Biologically-inspired simulations may support the study of genomic variation mechanisms in pathogens whose genomes have been recently sequenced.