Bayesian network approach to cell signaling pathway modeling

Dynamic label-free analysis of SARS-CoV-2 infection reveals virus-induced subcellular remodeling

Assessing the impact of SARS-CoV-2 on organelle dynamics allows a better understanding of the mechanisms of viral replication. We combine label-free holotomographic microscopy with Artificial Intelligence to visualize and quantify the subcellular changes triggered by SARS-CoV-2 infection. We study the dynamics of shape, position and dry mass of nucleoli, nuclei, lipid droplets and mitochondria within hundreds of single cells from early infection to syncytia formation and death. SARS-CoV-2 infection enlarges nucleoli, perturbs lipid droplets, changes mitochondrial shape and dry mass, and separates lipid droplets from mitochondria. We then used Bayesian network modeling on organelle dry mass states to define organelle cross-regulation networks and report modifications of organelle cross-regulation that are triggered by infection and syncytia formation. Our work highlights the subcellular remodeling induced by SARS-CoV-2 infection and provides an Artificial Intelligence-enhanced, label-free methodology to study in real-time the dynamics of cell populations and their content.

Fluorescent biosensor imaging meets deterministic mathematical modelling: quantitative investigation of signalling compartmentalization

Cells execute specific responses to diverse environmental cues by encoding information in distinctly compartmentalized biochemical signalling reactions. Genetically encoded fluorescent biosensors enable the spatial and temporal monitoring of signalling events in live cells. Temporal and spatiotemporal computational models can be used to interpret biosensor experiments in complex biochemical networks and to explore hypotheses that are difficult to test experimentally. In this review, we first provide brief discussions of the experimental toolkit of fluorescent biosensors as well as computational basics with a focus on temporal and spatiotemporal deterministic models. We then describe how we used this combined approach to identify and investigate a protein kinase A (PKA) - cAMP - Ca2+ oscillatory circuit in MIN6 β cells, a mouse pancreatic β cell system. We describe the application of this combined approach to interrogate how this oscillatory circuit is differentially regulated in a nano-compartment formed at the plasma membrane by the scaffolding protein A kinase anchoring protein 79/150. We leveraged both temporal and spatiotemporal deterministic models to identify the key regulators of this oscillatory circuit, which we confirmed with further experiments. The powerful approach of combining live-cell biosensor imaging with quantitative modelling, as discussed here, should find widespread use in the investigation of spatiotemporal regulation of cell signalling.

Data-driven learning how oncogenic gene expression locally alters heterocellular networks

Developing drugs increasingly relies on mechanistic modeling and simulation. Models that capture causal relations among genetic drivers of oncogenesis, functional plasticity, and host immunity complement wet experiments. Unfortunately, formulating such mechanistic cell-level models currently relies on hand curation, which can bias how data is interpreted or the priority of drug targets. In modeling molecular-level networks, rules and algorithms are employed to limit a priori biases in formulating mechanistic models. Here we combine digital cytometry with Bayesian network inference to generate causal models of cell-level networks linking an increase in gene expression associated with oncogenesis with alterations in stromal and immune cell subsets from bulk transcriptomic datasets. We predict how increased Cell Communication Network factor 4, a secreted matricellular protein, alters the tumor microenvironment using data from patients diagnosed with breast cancer and melanoma. Predictions are then tested using two immunocompetent mouse models for melanoma, which provide consistent experimental results.

While mechanistic models play increasing roles in immuno-oncology, hand network curation is current practice. Here the authors use a Bayesian data-driven approach to infer how expression of a secreted oncogene alters the cellular landscape within the tumor.

Bayesian networks elucidate complex genomic landscapes in cancer

Bayesian networks (BNs) are disciplined, explainable Artificial Intelligence models that can describe structured joint probability spaces. In the context of understanding complex relations between a number of variables in biological settings, they can be constructed from observed data and can provide a guiding, graphical tool in exploring such relations. Here we propose BNs for elucidating the relations between driver events in large cancer genomic datasets. We present a methodology that is specifically tailored to biologists and clinicians as they are the main producers of such datasets. We achieve this by using an optimal BN learning algorithm based on well established likelihood functions and by utilising just two tuning parameters, both of which are easy to set and have intuitive readings. To enhance value to clinicians, we introduce (a) the use of heatmaps for families in each network, and (b) visualising pairwise co-occurrence statistics on the network. For binary data, an optional step of fitting logic gates can be employed. We show how our methodology enhances pairwise testing and how biologists and clinicians can use BNs for discussing the main relations among driver events in large genomic cohorts. We demonstrate the utility of our methodology by applying it to 5 cancer datasets revealing complex genomic landscapes. Our networks identify central patterns in all datasets including a central 4-way mutual exclusivity between HDR, t(4,14), t(11,14) and t(14,16) in myeloma, and a 3-way mutual exclusivity of three major players: CALR, JAK2 and MPL, in myeloproliferative neoplasms. These analyses demonstrate that our methodology can play a central role in the study of large genomic cancer datasets.

Bayesian network inference on several blood and solid cancer genomic datasets provides more accessible ways to explore driver events in cancer.

The active kinome: The modern view of how active protein kinase networks fit in biological research

Biological regulatory networks are dynamic, intertwined, and complex systems making them challenging to study. While quantitative measurements of transcripts and proteins are key to investigate the state of a biological system, they do not inform the "active" state of regulatory networks. In consideration of that fact, "functional" proteomics assessments are needed to decipher active regulatory processes. Phosphorylation, a key post-translation modification, is a reversible regulatory mechanism that controls the functional state of proteins. Recent advancements of high-throughput protein kinase activity profiling platforms allow for a broad assessment of protein kinase networks in complex biological systems. In conjunction with sophisticated computational modeling techniques, these profiling platforms provide datasets that inform the active state of regulatory systems in disease models and highlight potential drug targets. Taken together, system-wide profiling of protein kinase activity has become a critical component of modern molecular biology research and presents a promising avenue for drug discovery.

Impact of explicit area scaling on kinetic models involving multiple compartments

BACKGROUND

Computational modelling of cell biological processes is a frequently used technique to analyse the underlying mechanisms and to generally understand the behaviour of these processes in the context of a pathway, network or even the whole cell. The most common technique in this context is the usage of ordinary differential equations that describe the kinetics of the relevant processes in mechanistic detail. Here, it is usually assumed that the content of the cell is well-stirred and thus homogeneous - which is of course an over-simplification, but often worked in the past. However, many processes happen at membranes and thus not in 3D, but in 2D. The scaling of the rates of these processes poses a special problem, if volumes of compartments are changed. They will typically scale with an area, but not with the volume of the involved compartment. However, commonly, this is neglected when setting up models and/or volume scaling also sometimes automatically happens when using modelling software in the field.

RESULTS

Here, we investigate generic as well as specific, realistic cases to find out, how strong the impact of the wrong scaling is for the outcome of simulations. We show that the importance of correct area scaling depends on the architecture of the reaction site and its changes upon volume alterations and it is hard to foresee, if it has a significant impact or not just by looking at the original model set-up. Moreover, scaled rates might exhibit more or less control over the behaviour of the system and therefore, accordingly, incorrect scaling will have more or less influence.

CONCLUSIONS

Working with multi-compartment reactions requires a careful consideration of the correct scaling of the rates when changing the volumes of the involved compartments. The error following incorrect scaling - often done by scaling with the volume of the respective compartments can lead to significant aberrations of model behaviour.

A unifying modelling formalism for the integration of stoichiometric and kinetic models

Current research on systems and synthetic biology relies heavily on mathematical models of the systems under study. The usefulness of such models depends on the quantity and quality of biological data, and on the availability of appropriate modelling formalisms that can gather and accommodate such data so that they can be exploited properly. Given our incomplete knowledge of biological systems and the fact that they consist of many subsystems, biological data are usually uncertain and heterogeneous. These facts hinder the use of mathematical models and computational methods. In the scope of dynamic biological systems, e.g. metabolic networks, this difficulty can be overcome by the novel modelling formalism of flexible nets (FNs). We show that an FN can combine, in a natural way, a stoichiometric model and a kinetic model. Moreover, the resulting net admits nonlinear dynamics and can be analysed in both transient and steady states.

Exploiting network topology for large-scale inference of nonlinear reaction models

The development of chemical reaction models aids understanding and prediction in areas ranging from biology to electrochemistry and combustion. A systematic approach to building reaction network models uses observational data not only to estimate unknown parameters but also to learn model structure. Bayesian inference provides a natural approach to this data-driven construction of models. Yet traditional Bayesian model inference methodologies that numerically evaluate the evidence for each model are often infeasible for nonlinear reaction network inference, as the number of plausible models can be combinatorially large. Alternative approaches based on model-space sampling can enable large-scale network inference, but their realization presents many challenges. In this paper, we present new computational methods that make large-scale nonlinear network inference tractable. First, we exploit the topology of networks describing potential interactions among chemical species to design improved 'between-model' proposals for reversible-jump Markov chain Monte Carlo. Second, we introduce a sensitivity-based determination of move types which, when combined with network-aware proposals, yields significant additional gains in sampling performance. These algorithms are demonstrated on inference problems drawn from systems biology, with nonlinear differential equation models of species interactions.

Systems biology primer: the basic methods and approaches

Systems biology is an integrative discipline connecting the molecular components within a single biological scale and also among different scales (e.g. cells, tissues and organ systems) to physiological functions and organismal phenotypes through quantitative reasoning, computational models and high-throughput experimental technologies. Systems biology uses a wide range of quantitative experimental and computational methodologies to decode information flow from genes, proteins and other subcellular components of signaling, regulatory and functional pathways to control cell, tissue, organ and organismal level functions. The computational methods used in systems biology provide systems-level insights to understand interactions and dynamics at various scales, within cells, tissues, organs and organisms. In recent years, the systems biology framework has enabled research in quantitative and systems pharmacology and precision medicine for complex diseases. Here, we present a brief overview of current experimental and computational methods used in systems biology.

Highly multiplexed immunofluorescence imaging of human tissues and tumors using t-CyCIF and conventional optical microscopes

The architecture of normal and diseased tissues strongly influences the development and progression of disease as well as responsiveness and resistance to therapy. We describe a tissue-based cyclic immunofluorescence (t-CyCIF) method for highly multiplexed immuno-fluorescence imaging of formalin-fixed, paraffin-embedded (FFPE) specimens mounted on glass slides, the most widely used specimens for histopathological diagnosis of cancer and other diseases. t-CyCIF generates up to 60-plex images using an iterative process (a cycle) in which conventional low-plex fluorescence images are repeatedly collected from the same sample and then assembled into a high-dimensional representation. t-CyCIF requires no specialized instruments or reagents and is compatible with super-resolution imaging; we demonstrate its application to quantifying signal transduction cascades, tumor antigens and immune markers in diverse tissues and tumors. The simplicity and adaptability of t-CyCIF makes it an effective method for pre-clinical and clinical research and a natural complement to single-cell genomics.

Unsupervised Network Analysis of the Plastic Supraoptic Nucleus Transcriptome Predicts Caprin2 Regulatory Interactions

The supraoptic nucleus (SON) is a group of neurons in the hypothalamus responsible for the synthesis and secretion of the peptide hormones vasopressin and oxytocin. Following physiological cues, such as dehydration, salt-loading and lactation, the SON undergoes a function related plasticity that we have previously described in the rat at the transcriptome level. Using the unsupervised graphical lasso (Glasso) algorithm, we reconstructed a putative network from 500 plastic SON genes in which genes are the nodes and the edges are the inferred interactions. The most active nodal gene identified within the network was Caprin2. Caprin2 encodes an RNA-binding protein that we have previously shown to be vital for the functioning of osmoregulatory neuroendocrine neurons in the SON of the rat hypothalamus. To test the validity of the Glasso network, we either overexpressed or knocked down Caprin2 transcripts in differentiated rat pheochromocytoma PC12 cells and showed that these manipulations had significant opposite effects on the levels of putative target mRNAs. These studies suggest that the predicative power of the Glasso algorithm within an in vivo system is accurate, and identifies biological targets that may be important to the functional plasticity of the SON.

Learning to read and write in evolution: from static pseudoenzymes and pseudosignalers to dynamic gear shifters

We present a systems biology view on pseudoenzymes that acknowledges that genes are not selfish: the genome is. With network function as the selectable unit, there has been an evolutionary bonus for recombination of functions of and within proteins. Many proteins house a functionality by which they 'read' the cell's state, and one by which they 'write' and thereby change that state. Should the writer domain lose its cognate function, a 'pseudoenzyme' or 'pseudosignaler' arises. GlnK involved in Escherichia coli ammonia assimilation may well be a pseudosignaler, associating 'reading' the nitrogen state of the cell to 'writing' the ammonium uptake activity. We identify functional pseudosignalers in the cyclin-dependent kinase complexes regulating cell-cycle progression. For the mitogen-activated protein kinase pathway, we illustrate how a 'dead' pseudosignaler could produce potentially selectable functionalities. Four billion years ago, bioenergetics may have shuffled 'electron-writers', producing various networks that all served the same function of anaerobic ATP synthesis and carbon assimilation from hydrogen and carbon dioxide, but at different ATP/acetate ratios. This would have enabled organisms to deal with variable challenges of energy need and substrate supply. The same principle might enable 'gear-shifting' in real time, by dynamically generating different pseudo-redox enzymes, reshuffling their coenzymes, and rerouting network fluxes. Non-stationary pH gradients in thermal vents together with similar such shuffling mechanisms may have produced a first selectable proton-motivated pyrophosphate synthase and subsequent ATP synthase. A combination of functionalities into enzymes, signalers, and the pseudo-versions thereof may offer fitness in terms of plasticity, both in real time and in evolution.

Enhanced construction of gene regulatory networks using hub gene information

BACKGROUND

Gene regulatory networks reveal how genes work together to carry out their biological functions. Reconstructions of gene networks from gene expression data greatly facilitate our understanding of underlying biological mechanisms and provide new opportunities for biomarker and drug discoveries. In gene networks, a gene that has many interactions with other genes is called a hub gene, which usually plays an essential role in gene regulation and biological processes. In this study, we developed a method for reconstructing gene networks using a partial correlation-based approach that incorporates prior information about hub genes. Through simulation studies and two real-data examples, we compare the performance in estimating the network structures between the existing methods and the proposed method.

RESULTS

In simulation studies, we show that the proposed strategy reduces errors in estimating network structures compared to the existing methods. When applied to Escherichia coli, the regulation network constructed by our proposed ESPACE method is more consistent with current biological knowledge than the SPACE method. Furthermore, application of the proposed method in lung cancer has identified hub genes whose mRNA expression predicts cancer progress and patient response to treatment.

CONCLUSIONS

We have demonstrated that incorporating hub gene information in estimating network structures can improve the performance of the existing methods.

A pedagogical walkthrough of computational modeling and simulation of Wnt signaling pathway using static causal models in MATLAB

Simulation study in systems biology involving computational experiments dealing with Wnt signaling pathways abound in literature but often lack a pedagogical perspective that might ease the understanding of beginner students and researchers in transition, who intend to work on the modeling of the pathway. This paucity might happen due to restrictive business policies which enforce an unwanted embargo on the sharing of important scientific knowledge. A tutorial introduction to computational modeling of Wnt signaling pathway in a human colorectal cancer dataset using static Bayesian network models is provided. The walkthrough might aid biologists/informaticians in understanding the design of computational experiments that is interleaved with exposition of the Matlab code and causal models from Bayesian network toolbox. The manuscript elucidates the coding contents of the advance article by Sinha (Integr. Biol. 6:1034-1048, 2014) and takes the reader in a step-by-step process of how (a) the collection and the transformation of the available biological information from literature is done, (b) the integration of the heterogeneous data and prior biological knowledge in the network is achieved, (c) the simulation study is designed, (d) the hypothesis regarding a biological phenomena is transformed into computational framework, and (e) results and inferences drawn using d-connectivity/separability are reported. The manuscript finally ends with a programming assignment to help the readers get hands-on experience of a perturbation project. Description of Matlab files is made available under GNU GPL v3 license at the Google code project on https://code.google.com/p/static-bn-for-wnt-signaling-pathway and https: //sites.google.com/site/shriprakashsinha/shriprakashsinha/projects/static-bn-for-wnt-signaling-pathway. Latest updates can be found in the latter website.

Identifying causal networks linking cancer processes and anti-tumor immunity using Bayesian network inference and metagene constructs

Cancer arises from a deregulation of both intracellular and intercellular networks that maintain system homeostasis. Identifying the architecture of these networks and how they are changed in cancer is a pre-requisite for designing drugs to restore homeostasis. Since intercellular networks only appear in intact systems, it is difficult to identify how these networks become altered in human cancer using many of the common experimental models. To overcome this, we used the diversity in normal and malignant human tissue samples from the Cancer Genome Atlas (TCGA) database of human breast cancer to identify the topology associated with intercellular networks in vivo. To improve the underlying biological signals, we constructed Bayesian networks using metagene constructs, which represented groups of genes that are concomitantly associated with different immune and cancer states. We also used bootstrap resampling to establish the significance associated with the inferred networks. In short, we found opposing relationships between cell proliferation and epithelial-to-mesenchymal transformation (EMT) with regards to macrophage polarization. These results were consistent across multiple carcinomas in that proliferation was associated with a type 1 cell-mediated anti-tumor immune response and EMT was associated with a pro-tumor anti-inflammatory response. To address the identifiability of these networks from other datasets, we could identify the relationship between EMT and macrophage polarization with fewer samples when the Bayesian network was generated from malignant samples alone. However, the relationship between proliferation and macrophage polarization was identified with fewer samples when the samples were taken from a combination of the normal and malignant samples. © 2016 American Institute of Chemical Engineers Biotechnol. Prog., 32:470-479, 2016.

Systematic analysis of BRAF(V600E) melanomas reveals a role for JNK/c-Jun pathway in adaptive resistance to drug-induced apoptosis

Drugs that inhibit RAF/MEK signaling, such as vemurafenib, elicit profound but often temporary anti-tumor responses in patients with BRAF^V600E melanoma. Adaptive responses to RAF/MEK inhibition occur on a timescale of hours to days, involve homeostatic responses that reactivate MAP kinase signaling and compensatory mitogenic pathways, and attenuate the anti-tumor effects of RAF/MEK inhibitors. We profile adaptive responses across a panel of melanoma cell lines using multiplex biochemical measurement, single-cell assays, and statistical modeling and show that adaptation involves at least six signaling cascades that act to reduce drug potency (IC₅₀) and maximal effect (i.e., E_max ≪ 1). Among these cascades, we identify a role for JNK/c-Jun signaling in vemurafenib adaptation and show that RAF and JNK inhibitors synergize in cell killing. This arises because JNK inhibition prevents a subset of cells in a cycling population from becoming quiescent upon vemurafenib treatment, thereby reducing drug E_max. Our findings demonstrate the breadth and diversity of adaptive responses to RAF/MEK inhibition and a means to identify which steps in a signaling cascade are most predictive of phenotypic response.

Methodologies for the modeling and simulation of biochemical networks, illustrated for signal transduction pathways: a primer

Biochemical networks depict the chemical interactions that take place among elements of living cells. They aim to elucidate how cellular behavior and functional properties of the cell emerge from the relationships between its components, i.e. molecules. Biochemical networks are largely characterized by dynamic behavior, and exhibit high degrees of complexity. Hence, the interest in such networks is growing and they have been the target of several recent modeling efforts. Signal transduction pathways (STPs) constitute a class of biochemical networks that receive, process, and respond to stimuli from the environment, as well as stimuli that are internal to the organism. An STP consists of a chain of intracellular signaling processes that ultimately result in generating different cellular responses. This primer presents the methodologies used for the modeling and simulation of biochemical networks, illustrated for STPs. These methodologies range from qualitative to quantitative, and include structural as well as dynamic analysis techniques. We describe the different methodologies, outline their underlying assumptions, and provide an assessment of their advantages and disadvantages. Moreover, publicly and/or commercially available implementations of these methodologies are listed as appropriate. In particular, this primer aims to provide a clear introduction and comprehensive coverage of biochemical modeling and simulation methodologies for the non-expert, with specific focus on relevant literature of STPs.

Inferring Aberrant Signal Transduction Pathways in Ovarian Cancer from TCGA Data

This paper concerns a new method for identifying aberrant signal transduction pathways (STPs) in cancer using case/control gene expression-level datasets, and applying that method and an existing method to an ovarian carcinoma dataset. Both methods identify STPs that are plausibly linked to all cancers based on current knowledge. Thus, the paper is most appropriate for the cancer informatics community. Our hypothesis is that STPs that are altered in tumorous tissue can be identified by applying a new Bayesian network (BN)-based method (causal analysis of STP aberration (CASA)) and an existing method (signaling pathway impact analysis (SPIA)) to the cancer genome atlas (TCGA) gene expression-level datasets. To test this hypothesis, we analyzed 20 cancer-related STPs and 6 randomly chosen STPs using the 591 cases in the TCGA ovarian carcinoma dataset, and the 102 controls in all 5 TCGA cancer datasets. We identified all the genes related to each of the 26 pathways, and developed separate gene expression datasets for each pathway. The results of the two methods were highly correlated. Furthermore, many of the STPs that ranked highest according to both methods are plausibly linked to all cancers based on current knowledge. Finally, CASA ranked the cancer-related STPs over the randomly selected STPs at a significance level below 0.05 (P = 0.047), but SPIA did not (P = 0.083).

Modeling signal transduction from protein phosphorylation to gene expression

BACKGROUND

Signaling networks are of great importance for us to understand the cell’s regulatory mechanism. The rise of large-scale genomic and proteomic data, and prior biological knowledge has paved the way for the reconstruction and discovery of novel signaling pathways in a data-driven manner. In this study, we investigate computational methods that integrate proteomics and transcriptomic data to identify signaling pathways transmitting signals in response to specific stimuli. Such methods can be applied to cancer genomic data to infer perturbed signaling pathways.

METHOD

We proposed a novel Bayesian Network (BN) framework to integrate transcriptomic data with proteomic data reflecting protein phosphorylation states for the purpose of identifying the pathways transmitting the signal of diverse stimuli in rat and human cells. We represented the proteins and genes as nodes in a BN in which edges reflect the regulatory relationship between signaling proteins. We designed an efficient inference algorithm that incorporated the prior knowledge of pathways and searched for a network structure in a data-driven manner.

RESULTS

We applied our method to infer rat and human specific networks given gene expression and proteomic datasets. We were able to effectively identify sparse signaling networks that modeled the observed transcriptomic and proteomic data. Our methods were able to identify distinct signaling pathways for rat and human cells in a data-driven manner, based on the facts that rat and human cells exhibited distinct transcriptomic and proteomics responses to a common set of stimuli. Our model performed well in the SBV IMPROVER challenge in comparison to other models addressing the same task. The capability of inferring signaling pathways in a data-driven fashion may contribute to cancer research by identifying distinct aberrations in signaling pathways underlying heterogeneous cancers subtypes.

Modeling the altered expression levels of genes on signaling pathways in tumors as causal bayesian networks

This paper concerns a study indicating that the expression levels of genes in signaling pathways can be modeled using a causal Bayesian network (BN) that is altered in tumorous tissue. These results open up promising areas of future research that can help identify driver genes and therapeutic targets. So, it is most appropriate for the cancer informatics community. Our central hypothesis is that the expression levels of genes that code for proteins on a signal transduction network (STP) are causally related and that this causal structure is altered when the STP is involved in cancer. To test this hypothesis, we analyzed 5 STPs associated with breast cancer, 7 STPs associated with other cancers, and 10 randomly chosen pathways, using a breast cancer gene expression level dataset containing 529 cases and 61 controls. We identified all the genes related to each of the 22 pathways and developed separate gene expression datasets for each pathway. We obtained significant results indicating that the causal structure of the expression levels of genes coding for proteins on STPs, which are believed to be implicated in both breast cancer and in all cancers, is more altered in the cases relative to the controls than the causal structure of the randomly chosen pathways.

Genome scale modeling in systems biology: algorithms and resources

In recent years, in silico studies and trial simulations have complemented experimental procedures. A model is a description of a system, and a system is any collection of interrelated objects; an object, moreover, is some elemental unit upon which observations can be made but whose internal structure either does not exist or is ignored. Therefore, any network analysis approach is critical for successful quantitative modeling of biological systems. This review highlights some of most popular and important modeling algorithms, tools, and emerging standards for representing, simulating and analyzing cellular networks in five sections. Also, we try to show these concepts by means of simple example and proper images and graphs. Overall, systems biology aims for a holistic description and understanding of biological processes by an integration of analytical experimental approaches along with synthetic computational models. In fact, biological networks have been developed as a platform for integrating information from high to low-throughput experiments for the analysis of biological systems. We provide an overview of all processes used in modeling and simulating biological networks in such a way that they can become easily understandable for researchers with both biological and mathematical backgrounds. Consequently, given the complexity of generated experimental data and cellular networks, it is no surprise that researchers have turned to computer simulation and the development of more theory-based approaches to augment and assist in the development of a fully quantitative understanding of cellular dynamics.

Cancer systems biology and modeling: microscopic scale and multiscale approaches

Cancer has become known as a complex and systematic disease on macroscopic, mesoscopic and microscopic scales. Systems biology employs state-of-the-art computational theories and high-throughput experimental data to model and simulate complex biological procedures such as cancer, which involves genetic and epigenetic, in addition to intracellular and extracellular complex interaction networks. In this paper, different systems biology modeling techniques such as systems of differential equations, stochastic methods, Boolean networks, Petri nets, cellular automata methods and agent-based systems are concisely discussed. We have compared the mentioned formalisms and tried to address the span of applicability they can bear on emerging cancer modeling and simulation approaches. Different scales of cancer modeling, namely, microscopic, mesoscopic and macroscopic scales are explained followed by an illustration of angiogenesis in microscopic scale of the cancer modeling. Then, the modeling of cancer cell proliferation and survival are examined on a microscopic scale and the modeling of multiscale tumor growth is explained along with its advantages.

Single timepoint models of dynamic systems

Many interesting studies aimed at elucidating the connectivity structure of biomolecular pathways make use of abundance measurements, and employ statistical and information theoretic approaches to assess connectivities. These studies often do not address the effects of the dynamics of the underlying biological system, yet dynamics give rise to impactful issues such as timepoint selection and its effect on structure recovery. In this work, we study conditions for reliable retrieval of the connectivity structure of a dynamic system, and the impact of dynamics on structure-learning efforts. We encounter an unexpected problem not previously described in elucidating connectivity structure from dynamic systems, show how this confounds structure learning of the system and discuss possible approaches to overcome the confounding effect. Finally, we test our hypotheses on an accurate dynamic model of the IGF signalling pathway. We use two structure-learning methods at four time points to contrast the performance and robustness of those methods in terms of recovering correct connectivity.

Engineered cell-based therapies: a vanguard of design-driven medicine

Engineered cell-based therapies are uniquely capable of performing sophisticated therapeutic functions in vivo, and this strategy is yielding promising clinical benefits for treating cancer. In this review, we discuss key opportunities and challenges for engineering customized cellular functions using cell-based therapy for cancer as a representative case study. We examine the historical development of chimeric antigen receptor (CAR) therapies as an illustration of the engineering design cycle. We also consider the potential roles that the complementary disciplines of systems biology and synthetic biology may play in realizing safe and effective treatments for a broad range of patients and diseases. In particular, we discuss how systems biology may facilitate both fundamental research and clinical translation, and we describe how the emerging field of synthetic biology is providing novel modalities for building customized cellular functions to overcome existing clinical barriers. Together, these approaches provide a powerful set of conceptual and experimental tools for transforming information into understanding, and for translating understanding into novel therapeutics to establish a new framework for design-driven medicine.

Boosting probabilistic graphical model inference by incorporating prior knowledge from multiple sources

Inferring regulatory networks from experimental data via probabilistic graphical models is a popular framework to gain insights into biological systems. However, the inherent noise in experimental data coupled with a limited sample size reduces the performance of network reverse engineering. Prior knowledge from existing sources of biological information can address this low signal to noise problem by biasing the network inference towards biologically plausible network structures. Although integrating various sources of information is desirable, their heterogeneous nature makes this task challenging. We propose two computational methods to incorporate various information sources into a probabilistic consensus structure prior to be used in graphical model inference. Our first model, called Latent Factor Model (LFM), assumes a high degree of correlation among external information sources and reconstructs a hidden variable as a common source in a Bayesian manner. The second model, a Noisy-OR, picks up the strongest support for an interaction among information sources in a probabilistic fashion. Our extensive computational studies on KEGG signaling pathways as well as on gene expression data from breast cancer and yeast heat shock response reveal that both approaches can significantly enhance the reconstruction accuracy of Bayesian Networks compared to other competing methods as well as to the situation without any prior. Our framework allows for using diverse information sources, like pathway databases, GO terms and protein domain data, etc. and is flexible enough to integrate new sources, if available.

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.

Optimization-based inference for temporally evolving networks with applications in biology

The problem of identifying dynamics of biological networks is of critical importance in order to understand biological systems. In this article, we propose a data-driven inference scheme to identify temporally evolving network representations of genetic networks. In the formulation of the optimization problem, we use an adjacency map as a priori information and define a cost function that both drives the connectivity of the graph to match biological data as well as generates a sparse and robust network at corresponding time intervals. Through simulation studies of simple examples, it is shown that this optimization scheme can help capture the topological change of a biological signaling pathway, and furthermore, might help to understand the structure and dynamics of biological genetic networks.

Oncogenic NRAS signaling differentially regulates survival and proliferation in melanoma

The discovery of potent inhibitors of the BRAF proto-oncogene has revolutionized therapy for melanoma harboring mutations in BRAF, yet NRAS-mutant melanoma remains without an effective therapy. Because direct pharmacological inhibition of the RAS proto-oncogene has thus far been unsuccessful, we explored systems biology approaches to identify synergistic drug combination(s) that can mimic RAS inhibition. Here, leveraging an inducible mouse model of NRAS-mutant melanoma, we show that pharmacological inhibition of mitogen-activated protein kinase kinase (MEK) activates apoptosis but not cell-cycle arrest, which is in contrast to complete genetic neuroblastoma RAS homolog (NRAS) extinction, which triggers both of these effects. Network modeling pinpointed cyclin-dependent kinase 4 (CDK4) as a key driver of this differential phenotype. Accordingly, combined pharmacological inhibition of MEK and CDK4 in vivo led to substantial synergy in therapeutic efficacy. We suggest a gradient model of oncogenic NRAS signaling in which the output is gated, resulting in the decoupling of discrete downstream biological phenotypes as a result of incomplete inhibition. Such a gated signaling model offers a new framework to identify nonobvious coextinction target(s) for combined pharmacological inhibition in NRAS-mutant melanomas.

Computational approaches for analyzing information flow in biological networks

The advancements in "omics" (proteomics, genomics, lipidomics, and metabolomics) technologies have yielded large inventories of genes, transcripts, proteins, and metabolites. The challenge is to find out how these entities work together to regulate the processes by which cells respond to external and internal signals. Mathematical and computational modeling of signaling networks has a key role in this task, and network analysis provides insights into biological systems and has applications for medicine. Here, we review experimental and theoretical progress and future challenges toward this goal. We focus on how networks are reconstructed from data, how these networks are structured to control the flow of biological information, and how the design features of the networks specify biological decisions.

Can case study approaches speed implementation of the NRC report: "toxicity testing in the 21st century: a vision and a strategy?"

The 2007 report "Toxicity Testing in the 21st Century: A Vision and a Strategy" argued for a change in toxicity testing for environmental agents and discussed federal funding mechanisms that could be used to support this transformation within the USA. The new approach would test for in vitro perturbations of toxicity pathways using human cells with high-throughput testing platforms. The NRC report proposed a deliberate timeline, spanning about 20 years, to implement a wholesale replacement of current in-life toxicity test approaches focused on apical responses with in vitro assays. One approach to accelerating implementation is to focus on well-studied prototype compounds with known toxicity pathway targets. Through a series of carefully executed case studies with four or five pathway prototypes, the various steps required for implementation of an in vitro toxicity pathway approach to risk assessment could be developed and refined. In this article, we discuss alternative approaches for implementation and also outline advantages of a case study approach and the manner in which the case studies could be pursued using current methodologies. A case study approach would be complementary to recently proposed efforts to map the human toxome, while representing a significant extension toward more formal risk assessment compared to the profiling and prioritization approaches inherent in programs such as the EPA's ToxCast effort.

Integrated inference and analysis of regulatory networks from multi-level measurements

Regulatory and signaling networks coordinate the enormously complex interactions and processes that control cellular processes (such as metabolism and cell division), coordinate response to the environment, and carry out multiple cell decisions (such as development and quorum sensing). Regulatory network inference is the process of inferring these networks, traditionally from microarray data but increasingly incorporating other measurement types such as proteomics, ChIP-seq, metabolomics, and mass cytometry. We discuss existing techniques for network inference. We review in detail our pipeline, which consists of an initial biclustering step, designed to estimate co-regulated groups; a network inference step, designed to select and parameterize likely regulatory models for the control of the co-regulated groups from the biclustering step; and a visualization and analysis step, designed to find and communicate key features of the network. Learning biological networks from even the most complete data sets is challenging; we argue that integrating new data types into the inference pipeline produces networks of increased accuracy, validity, and biological relevance.

A Bayesian Approach to Pathway Analysis by Integrating Gene-Gene Functional Directions and Microarray Data

Many statistical methods have been developed to screen for differentially expressed genes associated with specific phenotypes in the microarray data. However, it remains a major challenge to synthesize the observed expression patterns with abundant biological knowledge for more complete understanding of the biological functions among genes. Various methods including clustering analysis on genes, neural network, Bayesian network and pathway analysis have been developed toward this goal. In most of these procedures, the activation and inhibition relationships among genes have hardly been utilized in the modeling steps. We propose two novel Bayesian models to integrate the microarray data with the putative pathway structures obtained from the KEGG database and the directional gene-gene interactions in the medical literature. We define the symmetric Kullback-Leibler divergence of a pathway, and use it to identify the pathway(s) most supported by the microarray data. Monte Carlo Markov Chain sampling algorithm is given for posterior computation in the hierarchical model. The proposed method is shown to select the most supported pathway in an illustrative example. Finally, we apply the methodology to a real microarray data set to understand the gene expression profile of osteoblast lineage at defined stages of differentiation. We observe that our method correctly identifies the pathways that are reported to play essential roles in modulating bone mass.

Modeling formalisms in Systems Biology

Systems Biology has taken advantage of computational tools and high-throughput experimental data to model several biological processes. These include signaling, gene regulatory, and metabolic networks. However, most of these models are specific to each kind of network. Their interconnection demands a whole-cell modeling framework for a complete understanding of cellular systems. We describe the features required by an integrated framework for modeling, analyzing and simulating biological processes, and review several modeling formalisms that have been used in Systems Biology including Boolean networks, Bayesian networks, Petri nets, process algebras, constraint-based models, differential equations, rule-based models, interacting state machines, cellular automata, and agent-based models. We compare the features provided by different formalisms, and discuss recent approaches in the integration of these formalisms, as well as possible directions for the future.

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.

RegNetB: predicting relevant regulator-gene relationships in localized prostate tumor samples

Background

A central question in cancer biology is what changes cause a healthy cell to form a tumor. Gene expression data could provide insight into this question, but it is difficult to distinguish between a gene that causes a change in gene expression from a gene that is affected by this change. Furthermore, the proteins that regulate gene expression are often themselves not regulated at the transcriptional level. Here we propose a Bayesian modeling framework we term RegNetB that uses mechanistic information about the gene regulatory network to distinguish between factors that cause a change in expression and genes that are affected by the change. We test this framework using human gene expression data describing localized prostate cancer progression.

Results

The top regulatory relationships identified by RegNetB include the regulation of RLN1, RLN2, by PAX4, the regulation of ACPP (PAP) by JUN, BACH1 and BACH2, and the co-regulation of PGC and GDF15 by MAZ and TAF8. These target genes are known to participate in tumor progression, but the suggested regulatory roles of PAX4, BACH1, BACH2, MAZ and TAF8 in the process is new.

Conclusion

Integrating gene expression data and regulatory topologies can aid in identifying potentially causal mechanisms for observed changes in gene expression.

Partially observed bipartite network analysis to identify predictive connections in transcriptional regulatory networks

Background

Messenger RNA expression is regulated by a complex interplay of different regulatory proteins. Unfortunately, directly measuring the individual activity of these regulatory proteins is difficult, leaving us with only the resulting gene expression pattern as a marker for the underlying regulatory network or regulator-gene associations. Furthermore, traditional methods to predict these regulator-gene associations do not define the relative importance of each association, leading to a large number of connections in the global regulatory network that, although true, are not useful.

Results

Here we present a Bayesian method that identifies which known transcriptional relationships in a regulatory network are consistent with a given body of static gene expression data by eliminating the non-relevant ones. The Partially Observed Bipartite Network (POBN) approach developed here is tested using E. coli expression data and a transcriptional regulatory network derived from RegulonDB. When the regulatory network for E. coli was integrated with 266 E. coli gene chip observations, POBN identified 93 out of 570 connections that were either inconsistent or not adequately supported by the expression data.

Conclusion

POBN provides a systematic way to integrate known transcriptional networks with observed gene expression data to better identify which transcriptional pathways are likely responsible for the observed gene expression pattern.

From Physics to Pharmacology?

Over the last fifty years there has been an explosion of biological data, leading to the realization that to fully explain biological mechanisms it is necessary to interpret them as complex dynamical systems. The first stage of this interpretation is to determine which components (proteins, genes or metabolites) of the system interact. This is usually represented by a graph, or network. The behavior of this network can then be investigated using mathematical modeling. In vivo these biological networks show several remarkable (and seemingly paradoxical) properties including robustness, plasticity and sensitivity. Erroneous behavior of these networks is often associated with disease. Hence understanding the system-level properties can have important implications for the treatment of disease. Systems biology is an organized approach to quantitatively describe and elucidate the behavior of these complex networks. This review focuses on the progress and future challenges of a systems approach to biology.

From biological pathways to regulatory networks

This paper presents a general theoretical framework for generating Boolean networks whose state transitions realize a set of given biological pathways or minor variations thereof. This ill-posed inverse problem, which is of crucial importance across practically all areas of biology, is solved by using Karnaugh maps which are classical tools for digital system design. It is shown that the incorporation of prior knowledge, presented in the form of biological pathways, can bring about a dramatic reduction in the cardinality of the network search space. Constraining the connectivity of the network, the number and relative importance of the attractors, and concordance with observed time-course data are additional factors that can be used to further reduce the cardinality of the search space. The networks produced by the approaches developed here should facilitate the understanding of multivariate biological phenomena and the subsequent design of intervention approaches that are more likely to be successful in practice. As an example, the results of this paper are applied to the widely studied p53 pathway and it is shown that the resulting network exhibits dynamic behavior consistent with experimental observations from the published literature.

Task-oriented modular decomposition of biological networks: trigger mechanism in blood coagulation

Analysis of complex time-dependent biological networks is an important challenge in the current postgenomic era. We propose a middle-out approach for decomposition and analysis of complex time-dependent biological networks based on: 1), creation of a detailed mechanism-driven mathematical model of the network; 2), network response decomposition into several physiologically relevant subtasks; and 3), subsequent decomposition of the model, with the help of task-oriented necessity and sensitivity analysis into several modules that each control a single specific subtask, which is followed by further simplification employing temporal hierarchy reduction. The technique is tested and illustrated by studying blood coagulation. Five subtasks (threshold, triggering, control by blood flow velocity, spatial propagation, and localization), together with responsible modules, can be identified for the coagulation network. We show that the task of coagulation triggering is completely regulated by a two-step pathway containing a single positive feedback of factor V activation by thrombin. These theoretical predictions are experimentally confirmed by studies of fibrin generation in normal, factor V-, and factor VIII-deficient plasmas. The function of the factor V-dependent feedback is to minimize temporal and parametrical intervals of fibrin clot instability. We speculate that this pathway serves to lessen possibility of fibrin clot disruption by flow and subsequent thromboembolism.

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.

Simplistic pathways or complex networks?

Signaling events are frequently described in textbooks as linear cascades. However, in reality, input cues are processed by dynamic and context-specific networks, which are assembled from numerous signaling molecules. Diseases, such as cancer, are typically associated with multiple genomic alterations that likely change the structure and dynamics of cellular signaling networks. To assess the impact of such genomic alterations on the structure of signaling networks and on the ability of cells to accurately translate environmental cues into phenotypic changes, we argue studies must be conducted on a network level. Advances in technologies and computational approaches for data integration have permitted network studies of signaling events in both cancer and normal cells. Here we will review recent advances and how they have impacted our view on signaling networks with a specific angle on signal processing in cancer.

Multi-parameter phenotypic profiling: using cellular effects to characterize small-molecule compounds

Multi-parameter phenotypic profiling of small molecules provides important insights into their mechanisms of action, as well as a systems level understanding of biological pathways and their responses to small molecule treatments. It therefore deserves more attention at an early step in the drug discovery pipeline. Here, we summarize the technologies that are currently in use for phenotypic profiling--including mRNA-, protein- and imaging-based multi-parameter profiling--in the drug discovery context. We think that an earlier integration of phenotypic profiling technologies, combined with effective experimental and in silico target identification approaches, can improve success rates of lead selection and optimization in the drug discovery process.

Intelligently deciphering unintelligible designs: algorithmic algebraic model checking in systems biology

Systems biology, as a subject, has captured the imagination of both biologists and systems scientists alike. But what is it? This review provides one researcher's somewhat idiosyncratic view of the subject, but also aims to persuade young scientists to examine the possible evolution of this subject in a rich historical context. In particular, one may wish to read this review to envision a subject built out of a consilience of many interesting concepts from systems sciences, logic and model theory, and algebra, culminating in novel tools, techniques and theories that can reveal deep principles in biology--seen beyond mere observations. A particular focus in this review is on approaches embedded in an embryonic program, dubbed 'algorithmic algebraic model checking', and its powers and limitations.

Learning signaling network structures with sparsely distributed data

Flow cytometric measurement of signaling protein abundances has proved particularly useful for elucidation of signaling pathway structure. The single cell nature of the data ensures a very large dataset size, providing a statistically robust dataset for structure learning. Moreover, the approach is easily scaled to many conditions in high throughput. However, the technology suffers from a dimensionality constraint: at the cutting edge, only about 12 protein species can be measured per cell, far from sufficient for most signaling pathways. Because the structure learning algorithm (in practice) requires that all variables be measured together simultaneously, this restricts structure learning to the number of variables that constitute the flow cytometer's upper dimensionality limit. To address this problem, we present here an algorithm that enables structure learning for sparsely distributed data, allowing structure learning beyond the measurement technology's upper dimensionality limit for simultaneously measurable variables. The algorithm assesses pairwise (or n-wise) dependencies, constructs "Markov neighborhoods" for each variable based on these dependencies, measures each variable in the context of its neighborhood, and performs structure learning using a constrained search.

Learning cyclic signaling pathway structures while minimizing data requirements

Bayesian network structure learning is a useful tool for elucidation of regulatory structures of biomolecular pathways. The approach however is limited by its acyclicity constraint, a problematic one in the cycle-containing biological domain. Here, we introduce a novel method for modeling cyclic pathways in biology, by employing our newly introduced Generalized Bayesian Networks (GBNs). Our novel algorithm enables cyclic structure learning while employing biologically relevant data, as it extends our cycle-learning algorithm to permit learning with singly perturbed samples. We present theoretical arguments as well as structure learning results from realistic, simulated data of a biological system. We also present results from a real world dataset, involving signaling pathways in T-cells.

Biochemical basis for the functional switch that regulates hepatocyte growth factor receptor tyrosine kinase activation

Ligand-induced dimerization of receptor tyrosine kinases (RTKs) modulates a system of linked biochemical reactions, sharply switching the RTK from a quiescent state to an active state that becomes phosphorylated and triggers intracellular signaling pathways. To improve our understanding of this molecular switch, we developed a quantitative model for hepatocyte growth factor receptor (c-MET) activation using parameters derived in large part from c-MET kinetic and thermodynamic experiments. Our model accurately produces the qualitative and quantitative dynamic features of c-MET phosphorylation observed in cells following ligand binding, including a rapid transient buildup of phosphorylated c-MET at high ligand concentrations. In addition, our model predicts a slow buildup of phosphorylated c-MET under conditions of reduced phosphatase activity and no extracellular agonist. Significantly, this predicted response is observed in cells treated with phosphatase inhibitors, further validating our model. Parameter sensitivity studies clearly show that synergistic oligomerization-dependent changes in c-MET kinetic, thermodynamic, and dephosphorylation properties result in the selective activation of the dimeric receptor, confirming that this model can be used to accurately evaluate the relative importance of linked biochemical reactions important for c-MET activation. Our model suggests that the functional differences observed between c-MET monomers and dimers may have incrementally evolved to optimize cell surface signaling responses.

Negative autoregulation by FAS mediates robust fetal erythropoiesis

Tissue development is regulated by signaling networks that control developmental rate and determine ultimate tissue mass. Here we present a novel computational algorithm used to identify regulatory feedback and feedforward interactions between progenitors in developing erythroid tissue. The algorithm makes use of dynamic measurements of red cell progenitors between embryonic days 12 and 15 in the mouse. It selects for intercellular interactions that reproduce the erythroid developmental process and endow it with robustness to external perturbations. This analysis predicts that negative autoregulatory interactions arise between early erythroblasts of similar maturation stage. By studying embryos mutant for the death receptor FAS, or for its ligand, FASL, and by measuring the rate of FAS-mediated apoptosis in vivo, we show that FAS and FASL are pivotal negative regulators of fetal erythropoiesis, in the manner predicted by the computational model. We suggest that apoptosis in erythroid development mediates robust homeostasis regulating the number of red blood cells reaching maturity.

Non-linear dimensionality reduction of signaling networks

Background

Systems wide modeling and analysis of signaling networks is essential for understanding complex cellular behaviors, such as the biphasic responses to different combinations of cytokines and growth factors. For example, tumor necrosis factor (TNF) can act as a proapoptotic or prosurvival factor depending on its concentration, the current state of signaling network and the presence of other cytokines. To understand combinatorial regulation in such systems, new computational approaches are required that can take into account non-linear interactions in signaling networks and provide tools for clustering, visualization and predictive modeling.

Results

Here we extended and applied an unsupervised non-linear dimensionality reduction approach, Isomap, to find clusters of similar treatment conditions in two cell signaling networks: (I) apoptosis signaling network in human epithelial cancer cells treated with different combinations of TNF, epidermal growth factor (EGF) and insulin and (II) combination of signal transduction pathways stimulated by 21 different ligands based on AfCS double ligand screen data. For the analysis of the apoptosis signaling network we used the Cytokine compendium dataset where activity and concentration of 19 intracellular signaling molecules were measured to characterise apoptotic response to TNF, EGF and insulin. By projecting the original 19-dimensional space of intracellular signals into a low-dimensional space, Isomap was able to reconstruct clusters corresponding to different cytokine treatments that were identified with graph-based clustering. In comparison, Principal Component Analysis (PCA) and Partial Least Squares – Discriminant analysis (PLS-DA) were unable to find biologically meaningful clusters. We also showed that by using Isomap components for supervised classification with k-nearest neighbor (k-NN) and quadratic discriminant analysis (QDA), apoptosis intensity can be predicted for different combinations of TNF, EGF and insulin. Prediction accuracy was highest when early activation time points in the apoptosis signaling network were used to predict apoptosis rates at later time points. Extended Isomap also outperformed PCA on the AfCS double ligand screen data. Isomap identified more functionally coherent clusters than PCA and captured more information in the first two-components. The Isomap projection performs slightly worse when more signaling networks are analyzed; suggesting that the mapping function between cues and responses becomes increasingly non-linear when large signaling pathways are considered.

Conclusion

We developed and applied extended Isomap approach for the analysis of cell signaling networks. Potential biological applications of this method include characterization, visualization and clustering of different treatment conditions (i.e. low and high doses of TNF) in terms of changes in intracellular signaling they induce.