Flexible informatics for linking experimental data to mathematical models via DataRail

Modeling unveils sex differences of signaling networks in mouse embryonic stem cells

For a short period during early development of mammalian embryos, both X chromosomes in females are active, before dosage compensation is ensured through X-chromosome inactivation. In female mouse embryonic stem cells (mESCs), which carry two active X chromosomes, increased X-dosage affects cell signaling and impairs differentiation. The underlying mechanisms, however, remain poorly understood. To dissect X-dosage effects on the signaling network in mESCs, we combine systematic perturbation experiments with mathematical modeling. We quantify the response to a variety of inhibitors and growth factors for cells with one (XO) or two X chromosomes (XX). We then build models of the signaling networks in XX and XO cells through a semi-quantitative modeling approach based on modular response analysis. We identify a novel negative feedback in the PI3K/AKT pathway through GSK3. Moreover, the presence of a single active X makes mESCs more sensitive to the differentiation-promoting Activin A signal and leads to a stronger RAF1-mediated negative feedback in the FGF-triggered MAPK pathway. The differential response to these differentiation-promoting pathways can explain the impaired differentiation propensity of female mESCs.

kboolnet: a toolkit for the verification, validation, and visualization of reaction-contingency (rxncon) models

BACKGROUND

Computational models of cell signaling networks are extremely useful tools for the exploration of underlying system behavior and prediction of response to various perturbations. By representing signaling cascades as executable Boolean networks, the previously developed rxncon ("reaction-contingency") formalism and associated Python package enable accurate and scalable modeling of signal transduction even in large (thousands of components) biological systems. The models are split into reactions, which generate states, and contingencies, that impinge on reactions; this avoids the so-called "combinatorial explosion" of system size. Boolean description of the biological system compensates for the poor availability of kinetic parameters which are necessary for quantitative models. Unfortunately, few tools are available to support rxncon model development, especially for large, intricate systems.

RESULTS

We present the kboolnet toolkit ( https://github.com/Kufalab-UCSD/kboolnet , complete documentation at https://github.com/Kufalab-UCSD/kboolnet/wiki ), an R package and a set of scripts that seamlessly integrate with the python-based rxncon software and collectively provide a complete workflow for the verification, validation, and visualization of rxncon models. The verification script VerifyModel.R checks for responsiveness to repeated stimulations as well as consistency of steady state behavior. The validation scripts TruthTable.R, SensitivityAnalysis.R, and ScoreNet.R provide various readouts for the comparison of model predictions to experimental data. In particular, ScoreNet.R compares model predictions to a cloud-stored MIDAS-format experimental database to provide a numerical score for tracking model accuracy. Finally, the visualization scripts allow for graphical representations of model topology and behavior. The entire kboolnet toolkit is cloud-enabled, allowing for easy collaborative development; most scripts also allow for the extraction and analysis of individual user-defined "modules".

CONCLUSION

The kboolnet toolkit provides a modular, cloud-enabled workflow for the development of rxncon models, as well as their verification, validation, and visualization. This will enable the creation of larger, more comprehensive, and more rigorous models of cell signaling using the rxncon formalism in the future.

Mutations in ALK signaling pathways conferring resistance to ALK inhibitor treatment lead to collateral vulnerabilities in neuroblastoma cells

BACKGROUND

Development of resistance to targeted therapies has tempered initial optimism that precision oncology would improve poor outcomes for cancer patients. Resistance mechanisms, however, can also confer new resistance-specific vulnerabilities, termed collateral sensitivities. Here we investigated anaplastic lymphoma kinase (ALK) inhibitor resistance in neuroblastoma, a childhood cancer frequently affected by activating ALK alterations.

METHODS

Genome-wide forward genetic CRISPR-Cas9 based screens were performed to identify genes associated with ALK inhibitor resistance in neuroblastoma cell lines. Furthermore, the neuroblastoma cell line NBLW-R was rendered resistant by continuous exposure to ALK inhibitors. Genes identified to be associated with ALK inhibitor resistance were further investigated by generating suitable cell line models. In addition, tumor and liquid biopsy samples of four patients with ALK-mutated neuroblastomas before ALK inhibitor treatment and during tumor progression under treatment were genomically profiled.

RESULTS

Both genome-wide CRISPR-Cas9-based screens and preclinical spontaneous ALKi resistance models identified NF1 loss and activating NRASQ61K mutations to confer resistance to chemically diverse ALKi. Moreover, human neuroblastomas recurrently developed de novo loss of NF1 and activating RAS mutations after ALKi treatment, leading to therapy resistance. Pathway-specific perturbations confirmed that NF1 loss and activating RAS mutations lead to RAS-MAPK signaling even in the presence of ALKi. Intriguingly, NF1 loss rendered neuroblastoma cells hypersensitive to MEK inhibition.

CONCLUSIONS

Our results provide a clinically relevant mechanistic model of ALKi resistance in neuroblastoma and highlight new clinically actionable collateral sensitivities in resistant cells.

Tools for time-course simulation in systems biology: a brief overview

Dynamic modeling of biological systems is essential for understanding all properties of a given organism as it allows us to look not only at the static picture of an organism but also at its behavior under various conditions. With the increasing amount of experimental data, the number of tools that enable dynamic analysis also grows. However, various tools are based on different approaches, use different types of data and offer different functions for analyses; so it can be difficult to choose the most suitable tool for a selected type of model. Here, we bring a brief overview containing descriptions of 50 tools for the reconstruction of biological models, their time-course simulation and dynamic analysis. We examined each tool using test data and divided them based on the qualitative and quantitative nature of the mathematical apparatus they use.

Drug Effect Prediction by Integrating L1000 Genomic and Proteomic Big Data

The library of integrated Network-Based Cellular Signatures (LINCS) project aims to create a network-based understanding of biology by cataloging changes in gene expression and signal transduction. Gene expression and proteomic data in LINCS L1000 are cataloged for human cancer cells treated with compounds and genetic reagents. For understanding the related cell pathways and facilitating drug discovery, we developed binary linear programming (BLP) to infer cell-specific pathways and identify compounds' effects using L1000 gene expression and phosphoproteomics data. A generic pathway map for the MCF7 breast cancer cell line was built. Within them, BLP extracted the cell-specific pathways, which reliably predicted the compounds' effects. In this way, the potential drug effects are revealed by our models.

Systems Pharmacology Dissection of Cholesterol Regulation Reveals Determinants of Large Pharmacodynamic Variability between Cell Lines

In individuals, heterogeneous drug-response phenotypes result from a complex interplay of dose, drug specificity, genetic background, and environmental factors, thus challenging our understanding of the underlying processes and optimal use of drugs in the clinical setting. Here, we use mass-spectrometry-based quantification of molecular response phenotypes and logic modeling to explain drug-response differences in a panel of cell lines. We apply this approach to cellular cholesterol regulation, a biological process with high clinical relevance. From the quantified molecular phenotypes elicited by various targeted pharmacologic or genetic treatments, we generated cell-line-specific models that quantified the processes beneath the idiotypic intracellular drug responses. The models revealed that, in addition to drug uptake and metabolism, further cellular processes displayed significant pharmacodynamic response variability between the cell lines, resulting in cell-line-specific drug-response phenotypes. This study demonstrates the importance of integrating different types of quantitative systems-level molecular measurements with modeling to understand the effect of pharmacological perturbations on complex biological processes.

Logic Modeling in Quantitative Systems Pharmacology

Here we present logic modeling as an approach to understand deregulation of signal transduction in disease and to characterize a drug's mode of action. We discuss how to build a logic model from the literature and experimental data and how to analyze the resulting model to obtain insights of relevance for systems pharmacology. Our workflow uses the free tools OmniPath (network reconstruction from the literature), CellNOpt (model fit to experimental data), MaBoSS (model analysis), and Cytoscape (visualization).

Designing Drug-Response Experiments and Quantifying their Results

We developed a Python package to help in performing drug-response experiments at medium and high throughput and evaluating sensitivity metrics from the resulting data. In this article, we describe the steps involved in (1) generating files necessary for treating cells with the HP D300 drug dispenser, by pin transfer or by manual pipetting; (2) merging the data generated by high-throughput slide scanners, such as the Perkin Elmer Operetta, with treatment annotations; and (3) analyzing the results to obtain data normalized to untreated controls and sensitivity metrics such as IC₅₀ or GR₅₀ . These modules are available on GitHub and provide an automated pipeline for the design and analysis of high-throughput drug response experiments, that helps to prevent errors that can arise from manually processing large data files. © 2017 by John Wiley & Sons, Inc.

Data-driven reverse engineering of signaling pathways using ensembles of dynamic models

Despite significant efforts and remarkable progress, the inference of signaling networks from experimental data remains very challenging. The problem is particularly difficult when the objective is to obtain a dynamic model capable of predicting the effect of novel perturbations not considered during model training. The problem is ill-posed due to the nonlinear nature of these systems, the fact that only a fraction of the involved proteins and their post-translational modifications can be measured, and limitations on the technologies used for growing cells in vitro, perturbing them, and measuring their variations. As a consequence, there is a pervasive lack of identifiability. To overcome these issues, we present a methodology called SELDOM (enSEmbLe of Dynamic lOgic-based Models), which builds an ensemble of logic-based dynamic models, trains them to experimental data, and combines their individual simulations into an ensemble prediction. It also includes a model reduction step to prune spurious interactions and mitigate overfitting. SELDOM is a data-driven method, in the sense that it does not require any prior knowledge of the system: the interaction networks that act as scaffolds for the dynamic models are inferred from data using mutual information. We have tested SELDOM on a number of experimental and in silico signal transduction case-studies, including the recent HPN-DREAM breast cancer challenge. We found that its performance is highly competitive compared to state-of-the-art methods for the purpose of recovering network topology. More importantly, the utility of SELDOM goes beyond basic network inference (i.e. uncovering static interaction networks): it builds dynamic (based on ordinary differential equation) models, which can be used for mechanistic interpretations and reliable dynamic predictions in new experimental conditions (i.e. not used in the training). For this task, SELDOM's ensemble prediction is not only consistently better than predictions from individual models, but also often outperforms the state of the art represented by the methods used in the HPN-DREAM challenge.

Partial Least Squares Regression Models for the Analysis of Kinase Signaling

Partial least squares regression (PLSR) is a data-driven modeling approach that can be used to analyze multivariate relationships between kinase networks and cellular decisions or patient outcomes. In PLSR, a linear model relating an X matrix of dependent variables and a Y matrix of independent variables is generated by extracting the factors with the strongest covariation. While the identified relationship is correlative, PLSR models can be used to generate quantitative predictions for new conditions or perturbations to the network, allowing for mechanisms to be identified. This chapter will provide a brief explanation of PLSR and provide an instructive example to demonstrate the use of PLSR to analyze kinase signaling.

Context Specificity in Causal Signaling Networks Revealed by Phosphoprotein Profiling

Signaling networks downstream of receptor tyrosine kinases are among the most extensively studied biological networks, but new approaches are needed to elucidate causal relationships between network components and understand how such relationships are influenced by biological context and disease. Here, we investigate the context specificity of signaling networks within a causal conceptual framework using reverse-phase protein array time-course assays and network analysis approaches. We focus on a well-defined set of signaling proteins profiled under inhibition with five kinase inhibitors in 32 contexts: four breast cancer cell lines (MCF7, UACC812, BT20, and BT549) under eight stimulus conditions. The data, spanning multiple pathways and comprising ∼70,000 phosphoprotein and ∼260,000 protein measurements, provide a wealth of testable, context-specific hypotheses, several of which we experimentally validate. Furthermore, the data provide a unique resource for computational methods development, permitting empirical assessment of causal network learning in a complex, mammalian setting.

Time-course assays of signaling proteins in cancer cell lines under kinase inhibition

Causal conceptual framework for network analysis

Data shed light on causal protein networks that are specific to biological context

Resource for signaling biology and for benchmarking computational methods

Pathway and network analysis of cancer genomes

Genomic information on tumors from 50 cancer types cataloged by the International Cancer Genome Consortium (ICGC) shows that only a few well-studied driver genes are frequently mutated, in contrast to many infrequently mutated genes that may also contribute to tumor biology. Hence there has been large interest in developing pathway and network analysis methods that group genes and illuminate the processes involved. We provide an overview of these analysis techniques and show where they guide mechanistic and translational investigations.

Mixed-effects model of epithelial-mesenchymal transition reveals rewiring of signaling networks

The type II epithelial-mesenchymal transition (EMT) produces airway fibrosis and remodeling, contributing to the severity of asthma and chronic obstructive pulmonary disease. While numerous studies have been done on the mechanisms of the transition itself, few studies have investigated the system effects of EMT on signaling networks. Here, we use mixed effects modeling to develop a computational model of phospho-protein signaling data that compares human small airway epithelial cells (hSAECs) with their EMT-transformed counterparts across a series of perturbations with 8 ligands and 5 inhibitors, revealing previously uncharacterized changes in signaling in the EMT state. Strong couplings between menadione, TNFα and TGFβ and their known phospho-substrates were revealed after mixed effects modeling. Interestingly, the overall phospho-protein response was attenuated in EMT, with loss of Mena and TNFα coupling to heat shock protein (HSP)-27. These differences persisted after correction for EMT-induced changes in phospho-protein substrate abundance. Construction of network topology maps showed significant changes between the two cellular states, including a linkage between glycogen synthase kinase (GSK)-3α and small body size/mothers against decapentaplegic (SMAD)2. The model also predicted a loss of p38 mitogen activated protein kinase (p38MAPK)-independent HSP27 signaling, which we experimentally validated. We further characterized the relationship between HSP27 and signal transducers and activators of transcription (STAT)3 signaling, and determined that loss of HSP27 following EMT is only partially responsible for the downregulation of STAT3. These rewired connections represent therapeutic targets that could potentially reverse EMT and restore a normal phenotype to the respiratory mucosa.

Phosphoproteomic analyses reveal novel cross-modulation mechanisms between two signaling pathways in yeast

Cells respond to environmental stimuli via specialized signaling pathways. Concurrent stimuli trigger multiple pathways that integrate information, predominantly via protein phosphorylation. Budding yeast responds to NaCl and pheromone via two mitogen-activated protein kinase cascades, the high osmolarity, and the mating pathways, respectively. To investigate signal integration between these pathways, we quantified the time-resolved phosphorylation site dynamics after pathway co-stimulation. Using shotgun mass spectrometry, we quantified 2,536 phosphopeptides across 36 conditions. Our data indicate that NaCl and pheromone affect phosphorylation events within both pathways, which thus affect each other at more levels than anticipated, allowing for information exchange and signal integration. We observed a pheromone-induced down-regulation of Hog1 phosphorylation due to Gpd1, Ste20, Ptp2, Pbs2, and Ptc1. Distinct Ste20 and Pbs2 phosphosites responded differently to the two stimuli, suggesting these proteins as key mediators of the information exchange. A set of logic models was then used to assess the role of measured phosphopeptides in the crosstalk. Our results show that the integration of the response to different stimuli requires complex interconnections between signaling pathways.

In silico model-based inference: a contemporary approach for hypothesis testing in network biology

Inductive inference plays a central role in the study of biological systems where one aims to increase their understanding of the system by reasoning backwards from uncertain observations to identify causal relationships among components of the system. These causal relationships are postulated from prior knowledge as a hypothesis or simply a model. Experiments are designed to test the model. Inferential statistics are used to establish a level of confidence in how well our postulated model explains the acquired data. This iterative process, commonly referred to as the scientific method, either improves our confidence in a model or suggests that we revisit our prior knowledge to develop a new model. Advances in technology impact how we use prior knowledge and data to formulate models of biological networks and how we observe cellular behavior. However, the approach for model-based inference has remained largely unchanged since Fisher, Neyman and Pearson developed the ideas in the early 1900s that gave rise to what is now known as classical statistical hypothesis (model) testing. Here, I will summarize conventional methods for model-based inference and suggest a contemporary approach to aid in our quest to discover how cells dynamically interpret and transmit information for therapeutic aims that integrates ideas drawn from high performance computing, Bayesian statistics, and chemical kinetics.

RPPApipe: a pipeline for the analysis of reverse-phase protein array data

BACKGROUND AND SCOPE

Today, web-based data analysis pipelines exist for a wide variety of microarray platforms, such as ordinary gene-centered arrays, exon arrays and SNP arrays. However, most of the available software tools provide only limited support for reverse-phase protein arrays (RPPA), as relevant inherent properties of the corresponding datasets are not taken into account. Thus, we developed the web-based data analysis pipeline RPPApipe, which was specifically tailored to suit the characteristics of the RPPA platform and encompasses various tools for data preprocessing, statistical analysis, clustering and pathway analysis.

IMPLEMENTATION AND PERFORMANCE

All tools which are part of the RPPApipe software were implemented using R/Bioconductor. The software was embedded into our web-based ZBIT Bioinformatics Toolbox which is a customized instance of the Galaxy platform.

AVAILABILITY

RPPApipe is freely available under GNU Public License from http://webservices.cs.uni-tuebingen.de. A full documentation of the tool can be found on the corresponding website http://www.cogsys.cs.uni-tuebingen.de/software/RPPApipe.

The species translation challenge-a systems biology perspective on human and rat bronchial epithelial cells

The biological responses to external cues such as drugs, chemicals, viruses and hormones, is an essential question in biomedicine and in the field of toxicology, and cannot be easily studied in humans. Thus, biomedical research has continuously relied on animal models for studying the impact of these compounds and attempted to 'translate' the results to humans. In this context, the SBV IMPROVER (Systems Biology Verification for Industrial Methodology for PROcess VErification in Research) collaborative initiative, which uses crowd-sourcing techniques to address fundamental questions in systems biology, invited scientists to deploy their own computational methodologies to make predictions on species translatability. A multi-layer systems biology dataset was generated that was comprised of phosphoproteomics, transcriptomics and cytokine data derived from normal human (NHBE) and rat (NRBE) bronchial epithelial cells exposed in parallel to more than 50 different stimuli under identical conditions. The present manuscript describes in detail the experimental settings, generation, processing and quality control analysis of the multi-layer omics dataset accessible in public repositories for further intra- and inter-species translation studies.

Modeling of signaling pathways in chondrocytes based on phosphoproteomic and cytokine release data

OBJECTIVE

Chondrocyte signaling is widely identified as a key component in cartilage homeostasis. Dysregulations of the signaling processes in chondrocytes often result in degenerative diseases of the tissue. Traditionally, the literature has focused on the study of major players in chondrocyte signaling, but without considering the cross-talks between them. In this paper, we systematically interrogate the signal transduction pathways in chondrocytes, on both the phosphoproteomic and cytokine release levels.

METHODS

The signaling pathways downstream 78 receptors of interest are interrogated. On the phosphoproteomic level, 17 key phosphoproteins are measured upon stimulation with single treatments of 78 ligands. On the cytokine release level, 55 cytokines are measured in the supernatant upon stimulation with the same treatments. Using an Integer Linear Programming (ILP) formulation, the proteomic data is combined with a priori knowledge of proteins' connectivity to construct a mechanistic model, predictive of signal transduction in chondrocytes.

RESULTS

We were able to validate previous findings regarding major players of cartilage homeostasis and inflammation (e.g., IL1B, TNF, EGF, TGFA, INS, IGF1 and IL6). Moreover, we studied pro-inflammatory mediators (IL1B and TNF) together with pro-growth signals for investigating their role in chondrocytes hypertrophy and highlighted the role of underreported players such as Inhibin beta A (INHBA), Defensin beta 1 (DEFB1), CXCL1 and Flagellin, and uncovered the way they cross-react in the phosphoproteomic level.

CONCLUSIONS

The analysis presented herein, leveraged high throughput proteomic data via an ILP formulation to gain new insight into chondrocytes signaling and the pathophysiology of degenerative diseases in articular cartilage.

CoreFlow: a computational platform for integration, analysis and modeling of complex biological data

A major challenge in mass spectrometry and other large-scale applications is how to handle, integrate, and model the data that is produced. Given the speed at which technology advances and the need to keep pace with biological experiments, we designed a computational platform, CoreFlow, which provides programmers with a framework to manage data in real-time. It allows users to upload data into a relational database (MySQL), and to create custom scripts in high-level languages such as R, Python, or Perl for processing, correcting and modeling this data. CoreFlow organizes these scripts into project-specific pipelines, tracks interdependencies between related tasks, and enables the generation of summary reports as well as publication-quality images. As a result, the gap between experimental and computational components of a typical large-scale biology project is reduced, decreasing the time between data generation, analysis and manuscript writing. CoreFlow is being released to the scientific community as an open-sourced software package complete with proteomics-specific examples, which include corrections for incomplete isotopic labeling of peptides (SILAC) or arginine-to-proline conversion, and modeling of multiple/selected reaction monitoring (MRM/SRM) results.

BIOLOGICAL SIGNIFICANCE

CoreFlow was purposely designed as an environment for programmers to rapidly perform data analysis. These analyses are assembled into project-specific workflows that are readily shared with biologists to guide the next stages of experimentation. Its simple yet powerful interface provides a structure where scripts can be written and tested virtually simultaneously to shorten the life cycle of code development for a particular task. The scripts are exposed at every step so that a user can quickly see the relationships between the data, the assumptions that have been made, and the manipulations that have been performed. Since the scripts use commonly available programming languages, they can easily be transferred to and from other computational environments for debugging or faster processing. This focus on 'on the fly' analysis sets CoreFlow apart from other workflow applications that require wrapping of scripts into particular formats and development of specific user interfaces. Importantly, current and future releases of data analysis scripts in CoreFlow format will be of widespread benefit to the proteomics community, not only for uptake and use in individual labs, but to enable full scrutiny of all analysis steps, thus increasing experimental reproducibility and decreasing errors. This article is part of a Special Issue entitled: Can Proteomics Fill the Gap Between Genomics and Phenotypes?

Systematically studying kinase inhibitor induced signaling network signatures by integrating both therapeutic and side effects

Substantial effort in recent years has been devoted to analyzing data based large-scale biological networks, which provide valuable insight into the topologies of complex biological networks but are rarely context specific and cannot be used to predict the responses of cell signaling proteins to specific ligands or compounds. In this work, we proposed a novel strategy to investigate kinase inhibitor induced pathway signatures by integrating multiplex data in Library of Integrated Network-based Cellular Signatures (LINCS), e.g. KINOMEscan data and cell proliferation/mitosis imaging data. Using this strategy, we first established a PC9 cell line specific pathway model to investigate the pathway signatures in PC9 cell line when perturbed by a small molecule kinase inhibitor GW843682. This specific pathway revealed the role of PI3K/AKT in modulating the cell proliferation process and the absence of two anti-proliferation links, which indicated a potential mechanism of abnormal expansion in PC9 cell number. Incorporating the pathway model for side effects on primary human hepatocytes, it was used to screen 27 kinase inhibitors in LINCS database and PF02341066, known as Crizotinib, was finally suggested with an optimal concentration 4.6 uM to suppress PC9 cancer cell expansion while avoiding severe damage to primary human hepatocytes. Drug combination analysis revealed that the synergistic effect region can be predicted straightforwardly based on a threshold which is an inherent property of each kinase inhibitor. Furthermore, this integration strategy can be easily extended to other specific cell lines to be a powerful tool for drug screen before clinical trials.

Cyrface: An interface from Cytoscape to R that provides a user interface to R packages

There is an increasing number of software packages to analyse biological experimental data in the R environment. In particular, Bioconductor, a repository of curated R packages, is one of the most comprehensive resources for bioinformatics and biostatistics. The use of these packages is increasing, but it requires a basic understanding of the R language, as well as the syntax of the specific package used. The availability of user graphical interfaces for these packages would decrease the learning curve and broaden their application.   Here, we present a Cytoscape plug-in termed Cyrface that allows Cytoscape plug-ins to connect to any function and package developed in R. Cyrface can be used to run R packages from within the Cytoscape environment making use of a graphical user interface. Moreover, it links the R packages with the capabilities of Cytoscape and its plug-ins, in particular network visualization and analysis. Cyrface's utility has been demonstrated for two Bioconductor packages (CellNOptR and DrugVsDisease), and here we further illustrate its usage by implementing a workflow of data analysis and visualization. Download links, installation instructions and user guides can be accessed from the Cyrface homepage ( http://www.ebi.ac.uk/saezrodriguez/cyrface/).

Cyrface: An interface from Cytoscape to R that provides a user interface to R packages

There is an increasing number of software packages to analyse biological experimental data in the R environment. In particular, Bioconductor, a repository of curated R packages, is one of the most comprehensive resources for bioinformatics and biostatistics. The use of these packages is increasing, but it requires a basic understanding of the R language, as well as the syntax of the specific package used. The availability of user graphical interfaces for these packages would decrease the learning curve and broaden their application. Here, we present a Cytoscape app termed Cyrface that allows Cytoscape apps to connect to any function and package developed in R. Cyrface can be used to run R packages from within the Cytoscape environment making use of a graphical user interface. Moreover, it can link R packages with the capabilities of Cytoscape and its apps, in particular network visualization and analysis. Cyrface's utility has been demonstrated for two Bioconductor packages ( CellNOptR and DrugVsDisease), and here we further illustrate its usage by implementing a workflow of data analysis and visualization. Download links, installation instructions and user guides can be accessed from the Cyrface's homepage ( http://www.ebi.ac.uk/saezrodriguez/cyrface/) and from the Cytoscape app store ( http://apps.cytoscape.org/apps/cyrface).

Construction of cell type-specific logic models of signaling networks using CellNOpt

Mathematical models are useful tools for understanding protein signaling networks because they provide an integrated view of pharmacological and toxicological processes at the molecular level. Here we describe an approach previously introduced based on logic modeling to generate cell-specific, mechanistic and predictive models of signal transduction. Models are derived from a network encoding prior knowledge that is trained to signaling data, and can be either binary (based on Boolean logic) or quantitative (using a recently developed formalism, constrained fuzzy logic). The approach is implemented in the freely available tool CellNetOptimizer (CellNOpt). We explain the process CellNOpt uses to train a prior knowledge network to data and illustrate its application with a toy example as well as a realistic case describing signaling networks in the HepG2 liver cancer cell line.

Signaling network prediction by the Ontology Fingerprint enhanced Bayesian network

BACKGROUND

Despite large amounts of available genomic and proteomic data, predicting the structure and response of signaling networks is still a significant challenge. While statistical method such as Bayesian network has been explored to meet this challenge, employing existing biological knowledge for network prediction is difficult. The objective of this study is to develop a novel approach that integrates prior biological knowledge in the form of the Ontology Fingerprint to infer cell-type-specific signaling networks via data-driven Bayesian network learning; and to further use the trained model to predict cellular responses.

RESULTS

We applied our novel approach to address the Predictive Signaling Network Modeling challenge of the fourth (2009) Dialog for Reverse Engineering Assessment's and Methods (DREAM4) competition. The challenge results showed that our method accurately captured signal transduction of a network of protein kinases and phosphoproteins in that the predicted protein phosphorylation levels under all experimental conditions were highly correlated (R2 = 0.93) with the observed results. Based on the evaluation of the DREAM4 organizer, our team was ranked as one of the top five best performers in predicting network structure and protein phosphorylation activity under test conditions.

CONCLUSIONS

Bayesian network can be used to simulate the propagation of signals in cellular systems. Incorporating the Ontology Fingerprint as prior biological knowledge allows us to efficiently infer concise signaling network structure and to accurately predict cellular responses.

PROTEINCHALLENGE: crowd sourcing in proteomics analysis and software development

In large-scale proteomics studies there is a temptation, after months of experimental work, to plug resulting data into a convenient-if poorly implemented-set of tools, which may neither do the data justice nor help answer the scientific question. In this paper we have captured key concerns, including arguments for community-wide open source software development and "big data" compatible solutions for the future. For the meantime, we have laid out ten top tips for data processing. With these at hand, a first large-scale proteomics analysis hopefully becomes less daunting to navigate. However there is clearly a real need for robust tools, standard operating procedures and general acceptance of best practises. Thus we submit to the proteomics community a call for a community-wide open set of proteomics analysis challenges--PROTEINCHALLENGE--that directly target and compare data analysis workflows, with the aim of setting a community-driven gold standard for data handling, reporting and sharing.

Non Linear Programming (NLP) formulation for quantitative modeling of protein signal transduction pathways

Modeling of signal transduction pathways plays a major role in understanding cells' function and predicting cellular response. Mathematical formalisms based on a logic formalism are relatively simple but can describe how signals propagate from one protein to the next and have led to the construction of models that simulate the cells response to environmental or other perturbations. Constrained fuzzy logic was recently introduced to train models to cell specific data to result in quantitative pathway models of the specific cellular behavior. There are two major issues in this pathway optimization: i) excessive CPU time requirements and ii) loosely constrained optimization problem due to lack of data with respect to large signaling pathways. Herein, we address both issues: the former by reformulating the pathway optimization as a regular nonlinear optimization problem; and the latter by enhanced algorithms to pre/post-process the signaling network to remove parts that cannot be identified given the experimental conditions. As a case study, we tackle the construction of cell type specific pathways in normal and transformed hepatocytes using medium and large-scale functional phosphoproteomic datasets. The proposed Non Linear Programming (NLP) formulation allows for fast optimization of signaling topologies by combining the versatile nature of logic modeling with state of the art optimization algorithms.

CellNOptR: a flexible toolkit to train protein signaling networks to data using multiple logic formalisms

Background

Cells process signals using complex and dynamic networks. Studying how this is performed in a context and cell type specific way is essential to understand signaling both in physiological and diseased situations. Context-specific medium/high throughput proteomic data measured upon perturbation is now relatively easy to obtain but formalisms that can take advantage of these features to build models of signaling are still comparatively scarce.

Results

Here we present CellNOptR, an open-source R software package for building predictive logic models of signaling networks by training networks derived from prior knowledge to signaling (typically phosphoproteomic) data. CellNOptR features different logic formalisms, from Boolean models to differential equations, in a common framework. These different logic model representations accommodate state and time values with increasing levels of detail. We provide in addition an interface via Cytoscape (CytoCopteR) to facilitate use and integration with Cytoscape network-based capabilities.

Conclusions

Models generated with this pipeline have two key features. First, they are constrained by prior knowledge about the network but trained to data. They are therefore context and cell line specific, which results in enhanced predictive and mechanistic insights. Second, they can be built using different logic formalisms depending on the richness of the available data. Models built with CellNOptR are useful tools to understand how signals are processed by cells and how this is altered in disease. They can be used to predict the effect of perturbations (individual or in combinations), and potentially to engineer therapies that have differential effects/side effects depending on the cell type or context.

Sequential application of anticancer drugs enhances cell death by rewiring apoptotic signaling networks

Crosstalk and complexity within signaling pathways and their perturbation by oncogenes limit component-by-component approaches to understanding human disease. Network analysis of how normal and oncogenic signaling can be rewired by drugs may provide opportunities to target tumors with high specificity and efficacy. Using targeted inhibition of oncogenic signaling pathways, combined with DNA-damaging chemotherapy, we report that time-staggered EGFR inhibition, but not simultaneous coadministration, dramatically sensitizes a subset of triple-negative breast cancer cells to genotoxic drugs. Systems-level analysis-using high-density time-dependent measurements of signaling networks, gene expression profiles, and cell phenotypic responses in combination with mathematical modeling-revealed an approach for altering the intrinsic state of the cell through dynamic rewiring of oncogenic signaling pathways. This process converts these cells to a less tumorigenic state that is more susceptible to DNA damage-induced cell death by reactivation of an extrinsic apoptotic pathway whose function is suppressed in the oncogene-addicted state.

Nonparametric simulation of signal transduction networks with semi-synchronized update

Simulating signal transduction in cellular signaling networks provides predictions of network dynamics by quantifying the changes in concentration and activity-level of the individual proteins. Since numerical values of kinetic parameters might be difficult to obtain, it is imperative to develop non-parametric approaches that combine the connectivity of a network with the response of individual proteins to signals which travel through the network. The activity levels of signaling proteins computed through existing non-parametric modeling tools do not show significant correlations with the observed values in experimental results. In this work we developed a non-parametric computational framework to describe the profile of the evolving process and the time course of the proportion of active form of molecules in the signal transduction networks. The model is also capable of incorporating perturbations. The model was validated on four signaling networks showing that it can effectively uncover the activity levels and trends of response during signal transduction process.

An integrative model links multiple inputs and signaling pathways to the onset of DNA synthesis in hepatocytes

During liver regeneration, quiescent hepatocytes re-enter the cell cycle to proliferate and compensate for lost tissue. Multiple signals including hepatocyte growth factor, epidermal growth factor, tumor necrosis factor α, interleukin-6, insulin and transforming growth factor β orchestrate these responses and are integrated during the G₁ phase of the cell cycle. To investigate how these inputs influence DNA synthesis as a measure for proliferation, we established a large-scale integrated logical model connecting multiple signaling pathways and the cell cycle. We constructed our model based upon established literature knowledge, and successively improved and validated its structure using hepatocyte-specific literature as well as experimental DNA synthesis data. Model analyses showed that activation of the mitogen-activated protein kinase and phosphatidylinositol 3-kinase pathways was sufficient and necessary for triggering DNA synthesis. In addition, we identified key species in these pathways that mediate DNA replication. Our model predicted oncogenic mutations that were compared with the COSMIC database, and proposed intervention targets to block hepatocyte growth factor-induced DNA synthesis, which we validated experimentally. Our integrative approach demonstrates that, despite the complexity and size of the underlying interlaced network, logical modeling enables an integrative understanding of signaling-controlled proliferation at the cellular level, and thus can provide intervention strategies for distinct perturbation scenarios at various regulatory levels.

Construction of large signaling pathways using an adaptive perturbation approach with phosphoproteomic data

Construction of large and cell-specific signaling pathways is essential to understand information processing under normal and pathological conditions. On this front, gene-based approaches offer the advantage of large pathway exploration whereas phosphoproteomic approaches offer a more reliable view of pathway activities but are applicable to small pathway sizes. In this paper, we demonstrate an experimentally adaptive approach to construct large signaling pathways from phosphoproteomic data within a 3-day time frame. Our approach--taking advantage of the fast turnaround time of the xMAP technology--is carried out in four steps: (i) screen optimal pathway inducers, (ii) select the responsive ones, (iii) combine them in a combinatorial fashion to construct a phosphoproteomic dataset, and (iv) optimize a reduced generic pathway via an Integer Linear Programming formulation. As a case study, we uncover novel players and their corresponding pathways in primary human hepatocytes by interrogating the signal transduction downstream of 81 receptors of interest and constructing a detailed model for the responsive part of the network comprising 177 species (of which 14 are measured) and 365 interactions.

MCAM: multiple clustering analysis methodology for deriving hypotheses and insights from high-throughput proteomic datasets

Advances in proteomic technologies continue to substantially accelerate capability for generating experimental data on protein levels, states, and activities in biological samples. For example, studies on receptor tyrosine kinase signaling networks can now capture the phosphorylation state of hundreds to thousands of proteins across multiple conditions. However, little is known about the function of many of these protein modifications, or the enzymes responsible for modifying them. To address this challenge, we have developed an approach that enhances the power of clustering techniques to infer functional and regulatory meaning of protein states in cell signaling networks. We have created a new computational framework for applying clustering to biological data in order to overcome the typical dependence on specific a priori assumptions and expert knowledge concerning the technical aspects of clustering. Multiple clustering analysis methodology (‘MCAM’) employs an array of diverse data transformations, distance metrics, set sizes, and clustering algorithms, in a combinatorial fashion, to create a suite of clustering sets. These sets are then evaluated based on their ability to produce biological insights through statistical enrichment of metadata relating to knowledge concerning protein functions, kinase substrates, and sequence motifs. We applied MCAM to a set of dynamic phosphorylation measurements of the ERRB network to explore the relationships between algorithmic parameters and the biological meaning that could be inferred and report on interesting biological predictions. Further, we applied MCAM to multiple phosphoproteomic datasets for the ERBB network, which allowed us to compare independent and incomplete overlapping measurements of phosphorylation sites in the network. We report specific and global differences of the ERBB network stimulated with different ligands and with changes in HER2 expression. Overall, we offer MCAM as a broadly-applicable approach for analysis of proteomic data which may help increase the current understanding of molecular networks in a variety of biological problems.

Proteomic measurements, especially modification measurements, are greatly expanding the current knowledge of the state of proteins under various conditions. Harnessing these measurements to understand how these modifications are enzymatically regulated and their subsequent function in cellular signaling and physiology is a challenging new problem. Clustering has been very useful in reducing the dimensionality of many types of high-throughput biological data, as well inferring function of poorly understood molecular species. However, its implementation requires a great deal of technical expertise since there are a large number of parameters one must decide on in clustering, including data transforms, distance metrics, and algorithms. Previous knowledge of useful parameters does not exist for measurements of a new type. In this work we address two issues. First, we develop a framework that incorporates any number of possible parameters of clustering to produce a suite of clustering solutions. These solutions are then judged on their ability to infer biological information through statistical enrichment of existing biological annotations. Second, we apply this framework to dynamic phosphorylation measurements of the ERBB network, constructing the first extensive analysis of clustering of phosphoproteomic data and generating insight into novel components and novel functions of known components of the ERBB network.

Comparing signaling networks between normal and transformed hepatocytes using discrete logical models

Substantial effort in recent years has been devoted to constructing and analyzing large-scale gene and protein networks on the basis of "omic" data and literature mining. These interaction graphs provide valuable insight into the topologies of complex biological networks but are rarely context specific and cannot be used to predict the responses of cell signaling proteins to specific ligands or drugs. Conversely, traditional approaches to analyzing cell signaling are narrow in scope and cannot easily make use of network-level data. Here, we combine network analysis and functional experimentation by using a hybrid approach in which graphs are converted into simple mathematical models that can be trained against biochemical data. Specifically, we created Boolean logic models of immediate-early signaling in liver cells by training a literature-based prior knowledge network against biochemical data obtained from primary human hepatocytes and 4 hepatocellular carcinoma cell lines exposed to combinations of cytokines and small-molecule kinase inhibitors. Distinct families of models were recovered for each cell type, and these families clustered topologically into normal and diseased sets.

Combined logical and data-driven models for linking signalling pathways to cellular response

Background

Signalling pathways are the cornerstone on understanding cell function and predicting cell behavior. Recently, logical models of canonical pathways have been optimised with high-throughput phosphoproteomic data to construct cell-type specific pathways. However, less is known on how signalling pathways can be linked to a cellular response such as cell growth, death, cytokine secretion, or transcriptional activity.

Results

In this work, we measure the signalling activity (phosphorylation levels) and phenotypic behavior (cytokine secretion) of normal and cancer hepatocytes treated with a combination of cytokines and inhibitors. Using the two datasets, we construct "extended" pathways that integrate intracellular activity with cellular responses using a hybrid logical/data-driven computational approach. Boolean logic is used whenever a priori knowledge is accessible (i.e., construction of canonical pathways), whereas a data-driven approach is used for linking cellular behavior to signalling activity via non-canonical edges. The extended pathway is subsequently optimised to fit signalling and behavioural data using an Integer Linear Programming formulation. As a result, we are able to construct maps of primary and transformed hepatocytes downstream of 7 receptors that are capable of explaining the secretion of 22 cytokines.

Conclusions

We developed a method for constructing extended pathways that start at the receptor level and via a complex intracellular signalling pathway identify those mechanisms that drive cellular behaviour. Our results constitute a proof-of-principle for construction of "extended pathways" that are capable of linking pathway activity to diverse responses such as growth, death, differentiation, gene expression, or cytokine secretion.

Adaptive informatics for multifactorial and high-content biological data

Whereas genomic data are universally machine-readable, data arising from imaging, multiplex biochemistry, flow cytometry and other cell- and tissue-based assays usually reside in loosely organized files of poorly documented provenance. This arises because the relational databases used in genomic research are difficult to adapt to rapidly evolving experimental designs, data formats and analytic algorithms. Here we describe an adaptive approach to managing experimental data based on semantically-typed data hypercubes (SDCubes) that combine Hierarchical Data Format 5 (HDF5) and Extensible Markup Language (XML) file types. We demonstrate the application of SDCube-based storage using ImageRail, a software package for high-throughput microscopy. Experimental design and its day-to-day evolution, not rigid standards, determine how ImageRail data are organized in SDCubes. We apply ImageRail to the collection and analysis of drug dose-response landscapes in human cell lines at the single-cell level.

Training signaling pathway maps to biochemical data with constrained fuzzy logic: quantitative analysis of liver cell responses to inflammatory stimuli

Predictive understanding of cell signaling network operation based on general prior knowledge but consistent with empirical data in a specific environmental context is a current challenge in computational biology. Recent work has demonstrated that Boolean logic can be used to create context-specific network models by training proteomic pathway maps to dedicated biochemical data; however, the Boolean formalism is restricted to characterizing protein species as either fully active or inactive. To advance beyond this limitation, we propose a novel form of fuzzy logic sufficiently flexible to model quantitative data but also sufficiently simple to efficiently construct models by training pathway maps on dedicated experimental measurements. Our new approach, termed constrained fuzzy logic (cFL), converts a prior knowledge network (obtained from literature or interactome databases) into a computable model that describes graded values of protein activation across multiple pathways. We train a cFL-converted network to experimental data describing hepatocytic protein activation by inflammatory cytokines and demonstrate the application of the resultant trained models for three important purposes: (a) generating experimentally testable biological hypotheses concerning pathway crosstalk, (b) establishing capability for quantitative prediction of protein activity, and (c) prediction and understanding of the cytokine release phenotypic response. Our methodology systematically and quantitatively trains a protein pathway map summarizing curated literature to context-specific biochemical data. This process generates a computable model yielding successful prediction of new test data and offering biological insight into complex datasets that are difficult to fully analyze by intuition alone.

Interoperability of time series cytometric data: a cross platform approach for modeling tumor heterogeneity

The cell cycle, with its highly conserved features, is a fundamental driver for the temporal control of cell proliferation-while abnormal control and modulation of the cell cycle are characteristic of tumor cells. The principal aim in cancer biology is to seek an understanding of the origin and nature of innate and acquired heterogeneity at the cellular level, driven principally by temporal and functional asynchrony. A major bottleneck when mathematically modeling these biological systems is the lack of interlinked structured experimental data. This often results in the in silico models failing to translate the specific hypothesis into parameterized terms that enable robust validation and hence would produce suitable prediction tools rather than just simulation tools. The focus has been on linking data originating from different cytometric platforms and cell-based event analysis to inform and constrain the input parameters of a compartmental cell cycle model, hence partly measuring and deconvolving cell cycle heterogeneity within a tumor population. Our work has addressed the concept that the interoperability of cytometric data, derived from different cytometry platforms, can complement as well as enhance cellular parameters space, thus providing a more broader and in-depth view of the cellular systems. The initial aim was to enable the cell cycle model to deliver an improved integrated simulation of the well-defined and constrained biological system. From a modeling perspective, such a cross platform approach has provided a paradigm shift from conventional cross-validation approaches, and from a bioinformatics perspective, novel computational methodology has been introduced for integrating and mapping continuous data with cross-sectional data. This establishes the foundation for developing predictive models and in silico tracking and prediction of tumor progression

Setting the standards for signal transduction research

Major advances in high-throughput technology platforms, coupled with increasingly sophisticated computational methods for systematic data analysis, have provided scientists with tools to better understand the complexity of signaling networks. In this era of massive and diverse data collection, standardization efforts that streamline data gathering, analysis, storage, and sharing are becoming a necessity. Here, we give an overview of current technologies to study signal transduction. We argue that along with the opportunities the new technologies open, their heterogeneous nature poses critical challenges for data handling that are further increased when data are to be integrated in mathematical models. Efficient standardization through markup languages and data annotation is a sine qua non condition for a systems-level analysis of signaling processes. It remains to be seen the extent to which and the speed at which the emerging standardization efforts will be embraced by the signaling community.

Construction of signaling pathways and identification of drug effects on the liver cancer cell HepG2

Construction of signaling pathway maps and identification of drug effects are major challenge for pharmaceutical industries. Signaling maps are usually obtained from manual literature search, automated text mining algorithms, or canonical pathway databases (i.e. Reactome, KEGG, STKE, Pathway Studio, Ingenuity etc.) and in some cases they are used in combination with gene expression or mass spec data in an effort to create pathways specific to cell types or diseases. Our approach combines computational models with novel multicombinatorial high-throughput phosphoproteomic data for the functional analysis of signalling networks in mammalian cells. On the experimental front, we subject the cells with hundreds of co-treatment with a diverse set of ligands and inhibitors and we measure phosphorylation events on key signaling proteins using the xMAP technology. On the computational front, we create pathway maps that are cell type specific by fitting our phosphoprotein dataset into generic signaling maps via an Integer Linear programming formulation. To identify drug effects, we monitor the differences of topologies created with and without the presence of drug. In the present work, we use this approach to identify the effects of Nilotinib, a well known anti-cancer drug.

Elevated GM-CSF and IL-1beta levels compromise the ability of p38 MAPK inhibitors to modulate TNFalpha levels in the human monocytic/macrophage U937 cell line

Rheumatoid arthritis (RA) is a complex, multicellular disease involving a delicate balance between both pro- and anti-inflammatory cytokines which ultimately determines the disease phenotype. The simultaneous presence of multiple signaling molecules, and more specifically their relative levels, potentially influences the efficacy of directed therapies. Using the human U937 monocytic cell line, we generated a self-consistent dataset measuring 50 cytokines and 23 phosphoproteins in the presence of 6 small molecule inhibitors under 15 stimulatory conditions throughout a 24 hour time course. From this dataset, we are able to explore phosphoprotein and cytokine relationships, as well as evaluate the significance of cellular context on the ability of small molecule inhibitors to block inflammatory processes. We show that the ability of a p38 inhibitor to attenuate TNFalpha production is influenced by local levels of GM-CSF and IL-1beta, two cytokines known to be elevated in the joints of RA patients. Within the cell, compensatory mechanisms between signaling pathways are apparent, as selective p38 MAPK inhibition results in the increased phosphorylation of other MAPKs (ERK and JNK) and their downstream substrates (CREB, c-Jun, and ATF-2). Further, we demonstrate that TNFalpha-neutralizing antibodies have secondary effects on cytokine production, impacting more than just TNFalpha alone. p38 MAPK inhibition using a small molecule inhibitor also blocks production of anti-inflammatory cytokines including IL-10, IL-1ra and IL-2ra. Collectively, the impact of cell context on TNFalpha production and unintended blockade of anti-inflammatory cytokines may compromise the efficacy of p38 inhibitors in a clinical setting. The effort described in this work evaluates the effect of inhibitors on multiple endpoints (both intra- and extracellular), under a range of biologically relevant conditions, thus providing a unique means for differentiation of compounds and potential opportunity for improved pharmacological manipulation of disease endpoints in RA.

Cytokine-associated drug toxicity in human hepatocytes is associated with signaling network dysregulation

Idiosyncratic drug hepatotoxicity is a major problem in pharmaceutical development due to poor prediction capability of standard preclinical toxicity assessments and limited knowledge of its underlying mechanisms. Findings in animal models have shown that adverse effects of numerous drugs with idiosyncratic hepatotoxicity in humans can be reproduced in the presence of coincident inflammatory cytokine signaling. Following these observations, we have recently developed an in vitro drug/inflammatory cytokine co-treatment approach that can reproduce clinical drug hepatotoxicity signatures-particularly for idiosyncratic drugs-in cultured primary human hepatocytes. These observations have suggested that drug-induced stresses may interact with cytokine signaling to induce hepatic cytotoxicity, but the hepatocyte signaling mechanisms governing these interactions are poorly understood. Here, we collect high-throughput phosphoprotein signaling and cytotoxicity measurements in cultured hepatocytes, from multiple human donors, treated with combinations of hepatotoxic drugs (e.g. trovafloxacin, clarithromycin) and cytokines (tumor necrosis factor-alpha, interferon-gamma, interleukin-1 alpha, and interleukin-6). We demonstrate, through orthogonal partial least-squares regression (OPLSR) modeling of these signal-response data, that drug/cytokine hepatic cytotoxicity is integratively controlled by four key signaling pathways: Akt, p70 S6 kinase, MEK-ERK, and p38-HSP27. This modeling predicted, and experimental studies confirmed, that the MEK-ERK and p38-HSP27 pathways contribute pro-death signaling influences in drug/cytokine hepatic cytotoxicity synergy. Further, our four-pathway OPLSR model produced successful prediction of drug/cytokine hepatic cytotoxicities across different human donors, even though signaling and cytotoxicity responses were both highly donor-specific. Our findings highlight the critical role of kinase signaling in drug/cytokine hepatic cytotoxicity synergies and reveal that hepatic cytotoxicity responses are governed by multi-pathway signaling network balance.

Networks inferred from biochemical data reveal profound differences in toll-like receptor and inflammatory signaling between normal and transformed hepatocytes

Systematic study of cell signaling networks increasingly involves high throughput proteomics, transcriptional profiling, and automated literature mining with the aim of assembling large scale interaction networks. In contrast, functional analysis of cell signaling usually focuses on a much smaller sets of proteins and eschews computation but focuses directly on cellular responses to environment and perturbation. We sought to combine these two traditions by collecting cell response measures on a reasonably large scale and then attempting to infer differences in network topology between two cell types. Human hepatocytes and hepatocellular carcinoma cell lines were exposed to inducers of inflammation, innate immunity, and proliferation in the presence and absence of small molecule drugs, and multiplex biochemical measurement was then performed on intra- and extracellular signaling molecules. We uncovered major differences between primary and transformed hepatocytes with respect to the engagement of toll-like receptor and NF-κB-dependent secretion of chemokines and cytokines that prime and attract immune cells. Overall, our results serve as a proof of principle for an approach to network analysis that is systematic, comparative, and biochemically focused. More specifically, our data support the hypothesis that hepatocellular carcinoma cells down-regulate normal inflammatory and immune responses to avoid immune editing.

Towards a rigorous assessment of systems biology models: the DREAM3 challenges

Background

Systems biology has embraced computational modeling in response to the quantitative nature and increasing scale of contemporary data sets. The onslaught of data is accelerating as molecular profiling technology evolves. The Dialogue for Reverse Engineering Assessments and Methods (DREAM) is a community effort to catalyze discussion about the design, application, and assessment of systems biology models through annual reverse-engineering challenges.

Methodology and Principal Findings

We describe our assessments of the four challenges associated with the third DREAM conference which came to be known as the DREAM3 challenges: signaling cascade identification, signaling response prediction, gene expression prediction, and the DREAM3 in silico network challenge. The challenges, based on anonymized data sets, tested participants in network inference and prediction of measurements. Forty teams submitted 413 predicted networks and measurement test sets. Overall, a handful of best-performer teams were identified, while a majority of teams made predictions that were equivalent to random. Counterintuitively, combining the predictions of multiple teams (including the weaker teams) can in some cases improve predictive power beyond that of any single method.

Conclusions

DREAM provides valuable feedback to practitioners of systems biology modeling. Lessons learned from the predictions of the community provide much-needed context for interpreting claims of efficacy of algorithms described in the scientific literature.

Unveiling the role of network and systems biology in drug discovery

Network and systems biology offer a novel way of approaching drug discovery by developing models that consider the global physiological environment of protein targets, and the effects of modifying them, without losing the key molecular details. Here we review some recent advances in network and systems biology applied to human health, and discuss how they can have a big impact on some of the most interesting areas of drug discovery. In particular, we claim that network biology will play a central part in the development of novel polypharmacology strategies to fight complex multifactorial diseases, where efficacious therapies will need to center on altering entire pathways rather than single proteins. We briefly present new developments in the two areas where we believe network and system biology strategies are more likely to have an immediate contribution: predictive toxicology and drug repurposing.

DataPflex: a MATLAB-based tool for the manipulation and visualization of multidimensional datasets

MOTIVATION

DataPflex is a MATLAB-based application that facilitates the manipulation and visualization of multidimensional datasets. The strength of DataPflex lies in the intuitive graphical user interface for the efficient incorporation, manipulation and visualization of high-dimensional data that can be generated by multiplexed protein measurement platforms including, but not limited to Luminex or Meso-Scale Discovery. Such data can generally be represented in the form of multidimensional datasets [for example (time x stimulation x inhibitor x inhibitor concentration x cell type x measurement)]. For cases where measurements are made in a combinational fashion across multiple dimensions, there is a need for a tool to efficiently manipulate and reorganize such data for visualization. DataPflex accepts data consisting of up to five arbitrary dimensions in addition to a measurement dimension. Data are imported from a simple .xls format and can be exported to MATLAB or .xls. Data dimensions can be reordered, subdivided, merged, normalized and visualized in the form of collections of line graphs, bar graphs, surface plots, heatmaps, IC50's and other custom plots. Open source implementation in MATLAB enables easy extension for custom plotting routines and integration with more sophisticated analysis tools.

AVAILABILITY

DataPflex is distributed under the GPL license (http://www.gnu.org/licenses/) together with documentation, source code and sample data files at: http://code.google.com/p/datapflex.

SUPPLEMENTARY INFORMATION

Supplementary data available at Bioinformatics online.

Discrete logic modelling as a means to link protein signalling networks with functional analysis of mammalian signal transduction

Large-scale protein signalling networks are useful for exploring complex biochemical pathways but do not reveal how pathways respond to specific stimuli. Such specificity is critical for understanding disease and designing drugs. Here we describe a computational approach—implemented in the free CNO software—for turning signalling networks into logical models and calibrating the models against experimental data. When a literature-derived network of 82 proteins covering the immediate-early responses of human cells to seven cytokines was modelled, we found that training against experimental data dramatically increased predictive power, despite the crudeness of Boolean approximations, while significantly reducing the number of interactions. Thus, many interactions in literature-derived networks do not appear to be functional in the liver cells from which we collected our data. At the same time, CNO identified several new interactions that improved the match of model to data. Although missing from the starting network, these interactions have literature support. Our approach, therefore, represents a means to generate predictive, cell-type-specific models of mammalian signalling from generic protein signalling networks.

Identifying drug effects via pathway alterations using an integer linear programming optimization formulation on phosphoproteomic data

Understanding the mechanisms of cell function and drug action is a major endeavor in the pharmaceutical industry. Drug effects are governed by the intrinsic properties of the drug (i.e., selectivity and potency) and the specific signaling transduction network of the host (i.e., normal vs. diseased cells). Here, we describe an unbiased, phosphoproteomic-based approach to identify drug effects by monitoring drug-induced topology alterations. With our proposed method, drug effects are investigated under diverse stimulations of the signaling network. Starting with a generic pathway made of logical gates, we build a cell-type specific map by constraining it to fit 13 key phopshoprotein signals under 55 experimental conditions. Fitting is performed via an Integer Linear Program (ILP) formulation and solution by standard ILP solvers; a procedure that drastically outperforms previous fitting schemes. Then, knowing the cell's topology, we monitor the same key phosphoprotein signals under the presence of drug and we re-optimize the specific map to reveal drug-induced topology alterations. To prove our case, we make a topology for the hepatocytic cell-line HepG2 and we evaluate the effects of 4 drugs: 3 selective inhibitors for the Epidermal Growth Factor Receptor (EGFR) and a non-selective drug. We confirm effects easily predictable from the drugs' main target (i.e., EGFR inhibitors blocks the EGFR pathway) but we also uncover unanticipated effects due to either drug promiscuity or the cell's specific topology. An interesting finding is that the selective EGFR inhibitor Gefitinib inhibits signaling downstream the Interleukin-1alpha (IL1alpha) pathway; an effect that cannot be extracted from binding affinity-based approaches. Our method represents an unbiased approach to identify drug effects on small to medium size pathways which is scalable to larger topologies with any type of signaling interventions (small molecules, RNAi, etc). The method can reveal drug effects on pathways, the cornerstone for identifying mechanisms of drug's efficacy.