Temporal Ordering of Cancer Microarray Data through a Reinforcement Learning Based Approach

Temporal modeling and analysis and more specifically, temporal ordering are very important problems within the fields of bioinformatics and computational biology, as the temporal analysis of the events characterizing a certain biological process could provide significant insights into its development and progression. Particularly, in the case of cancer, understanding the dynamics and the evolution of this disease could lead to better methods for prediction and treatment. In this paper we tackle, from a computational perspective, the temporal ordering problem, which refers to constructing a sorted collection of multi-dimensional biological data, collection that reflects an accurate temporal evolution of biological systems. We introduce a novel approach, based on reinforcement learning, more precisely, on Q-learning, for the biological temporal ordering problem. The experimental evaluation is performed using several DNA microarray data sets, two of which contain cancer gene expression data. The obtained solutions are correlated either to the given correct ordering (in the cases where this is provided for validation), or to the overall survival time of the patients (in the case of the cancer data sets), thus confirming a good performance of the proposed model and indicating the potential of our proposal.     

An optimization framework of biological dynamical systems

Different biological dynamics are often described by different mathematical equations. On the other hand, some mathematical models describe many biological dynamics universally. Here, we focus on three biological dynamics: the Lotka-Volterra equation, the Hopfield neural networks, and the replicator equation. We describe these three dynamical models using a single optimization framework, which is constructed with employing the Riemannian geometry. Then, we show that the optimization structures of these dynamics are identical, and the differences among the three dynamics are only in the constraints of the optimization. From this perspective, we discuss the unified view for biological dynamics. We also discuss the plausible categorizations, the fundamental nature, and the efficient modeling of the biological dynamics, which arise from the optimization perspective of the dynamical systems.     

Hybrid Models and Biological Model Reduction with PyDSTool

The PyDSTool software environment is designed to develop, simulate, and analyze dynamical systems models, particularly for biological applications. Unlike the engineering application focus and graphical specification environments of most general purpose simulation tools, PyDSTool provides a programmatic environment well suited to exploratory data- and hypothesis-driven biological modeling problems. In this work, we show how the environment facilitates the application of hybrid dynamical modeling to the reverse engineering of complex biophysical dynamics; in this case, of an excitable membrane. The example demonstrates how the software provides novel tools that support the inference and validation of mechanistic hypotheses and the inclusion of data constraints in both quantitative and qualitative ways. The biophysical application is broadly relevant to models in the biosciences. The open source and platform-independent PyDSTool package is freely available under the BSD license from http://sourceforge.net/projects/pydstool/. The hosting service provides links to documentation and online forums for user support.     

Bioinformatics for the genomic sciences and towards systems biology. Japanese activities in the post-genome era

The knowledge gleaned from genome sequencing and post-genome analyses is having a very significant impact on a whole range of life sciences and their applications. 'Genome-wide analysis' is a good keyword to represent this tendency. Thanks to innovations in high-throughput measurement technologies and information technologies, genome-wide analysis is becoming available in a broad range of research fields from DNA sequences, gene and protein expressions, protein structures and interactions, to pathways or networks analysis. In fact, the number of research targets has increased by more than two orders in recent years and we should change drastically the attitude to research activities. The scope and speed of research activities are expanding and the field of bioinformatics is playing an important role. In parallel with the data-driven research approach that focuses on speedy handling and analyzing of the huge amount of data, a new approach is gradually gaining power. This is a 'model-driven research' approach, that incorporates biological modeling in its research framework. Computational simulations of biological processes play a pivotal role. By modeling and simulating, this approach aims at predicting and even designing the dynamic behaviors of complex biological systems, which is expected to make rapid progress in life science researches and lead to meaningful applications to various fields such as health care, food supply and improvement of environment. Genomic sciences are now advancing as great frontiers of research and applications in the 21st century.This article starts with surveying the general progress of bioinformatics (Section 1), and describes Japanese activities in bioinformatics (Section 2). In Section 3, I will introduce recent developments in Systems Biology which I think will become more important in the future.     

Information-Theoretic Analysis of the Dynamics of an Executable Biological Model

To facilitate analysis and understanding of biological systems, large-scale data are often integrated into models using a variety of mathematical and computational approaches. Such models describe the dynamics of the biological system and can be used to study the changes in the state of the system over time. For many model classes, such as discrete or continuous dynamical systems, there exist appropriate frameworks and tools for analyzing system dynamics. However, the heterogeneous information that encodes and bridges molecular and cellular dynamics, inherent to fine-grained molecular simulation models, presents significant challenges to the study of system dynamics. In this paper, we present an algorithmic information theory based approach for the analysis and interpretation of the dynamics of such executable models of biological systems. We apply a normalized compression distance (NCD) analysis to the state representations of a model that simulates the immune decision making and immune cell behavior. We show that this analysis successfully captures the essential information in the dynamics of the system, which results from a variety of events including proliferation, differentiation, or perturbations such as gene knock-outs. We demonstrate that this approach can be used for the analysis of executable models, regardless of the modeling framework, and for making experimentally quantifiable predictions.     

A Dynamic Workflow Approach for the Integration of Bioinformatics Services

Modern biological and chemical studies rely on life science databases as well as sophisticated software tools (e.g., homology search tools, modeling and visualization tools). These tools often have to be combined and integrated in order to support a given study. SIBIOS (System for the Integration of Bioinformatics Services) serves this purpose. The services are both life science database search services and software tools. The task engine is the core component of SIBIOS. It supports the execution of dynamic workflows that incorporate multiple bioinformatics services. The architecture of SIBIOS, the approaches to addressing the heterogeneity as well as interoperability of bioinformatics services, including data integration are presented in this paper.     

Superstability of the yeast cell-cycle dynamics: ensuring causality in the presence of biochemical stochasticity

Gene regulatory dynamics are governed by molecular processes and therefore exhibits an inherent stochasticity. However, for the survival of an organism it is a strict necessity that this intrinsic noise does not prevent robust functioning of the system. It is still an open question how dynamical stability is achieved in biological systems despite the omnipresent fluctuations. In this paper we investigate the cell cycle of the budding yeast Saccharomyces cerevisiae as an example of a well-studied organism. We study a genetic network model of 11 genes that coordinate the cell-cycle dynamics using a modeling framework which generalizes the concept of discrete threshold dynamics. By allowing for fluctuations in the process times, we introduce noise into the model, accounting for the effects of biochemical stochasticity. We study the dynamical attractor of the cell cycle and find a remarkable robustness against fluctuations of this kind. We identify mechanisms that ensure reliability in spite of fluctuations: 'Catcher states' and persistence of activity levels contribute significantly to the stability of the yeast cell cycle despite the inherent stochasticity.     

Swami一新一代生物学平台

生物信息学的发展产生了越来越多的数据库和生物学软件,研究人员在应用这些生物学工具处理实验数据时需要大量的时间解决数据格式转换和管理等问题。本文介绍了一种交互式的基于网络的新一代生物信息学分析平台一Swam,它综合了主要的生物信息学数据库和软件,可以加速数据处理并帮助用户管理数据。因此它将推动生物信息学向更深层次发展。     

Modeling of dynamical systems through deep learning

This review presents a modern perspective on dynamical systems in the context of current goals and open challenges. In particular, our review focuses on the key challenges of discovering dynamics from data and finding data-driven representations that make nonlinear systems amenable to linear analysis. We explore various challenges in modern dynamical systems, along with emerging techniques in data science and machine learning to tackle them. The two chief challenges are (1) nonlinear dynamics and (2) unknown or partially known dynamics. Machine learning is providing new and powerful techniques for both challenges. Dimensionality reduction methods are used for projecting dynamical methods in reduced form, and these methods perform computational efficiency on real-world data. Data-driven models drive to discover the governing equations and give laws of physics. The identification of dynamical systems through deep learning techniques succeeds in inferring physical systems. Machine learning provides advanced new and powerful algorithms for nonlinear dynamics. Advanced deep learning methods like autoencoders, recurrent neural networks, convolutional neural networks, and reinforcement learning are used in modeling of dynamical systems.     

Computational intelligence approaches for pattern discovery in biological systems

Biology, chemistry and medicine are faced by tremendous challenges caused by an overwhelming amount of data and the need for rapid interpretation. Computational intelligence (CI) approaches such as artificial neural networks, fuzzy systems and evolutionary computation are being used with increasing frequency to contend with this problem, in light of noise, non-linearity and temporal dynamics in the data. Such methods can be used to develop robust models of processes either on their own or in combination with standard statistical approaches. This is especially true for database mining, where modeling is a key component of scientific understanding. This review provides an introduction to current CI methods, their application to biological problems, and concludes with a commentary about the anticipated impact of these approaches in bioinformatics.     

Data-driven flow-map models for data-efficient discovery of dynamics and fast uncertainty quantification of biological and biochemical systems

Computational models are increasingly used to investigate and predict the complex dynamics of biological and biochemical systems. Nevertheless, governing equations of a biochemical system may not be (fully) known, which would necessitate learning the system dynamics directly from, often limited and noisy, observed data. On the other hand, when expensive models are available, systematic and efficient quantification of the effects of model uncertainties on quantities of interest can be an arduous task. This paper leverages the notion of flow-map (de)compositions to present a framework that can address both of these challenges via learning data-driven models useful for capturing the dynamical behavior of biochemical systems. Data-driven flow-map models seek to directly learn the integration operators of the governing differential equations in a black-box manner, irrespective of structure of the underlying equations. As such, they can serve as a flexible approach for deriving fast-to-evaluate surrogates for expensive computational models of system dynamics, or, alternatively, for reconstructing the long-term system dynamics via experimental observations. We present a data-efficient approach to data-driven flow-map modeling based on polynomial chaos Kriging. The approach is demonstrated for discovery of the dynamics of various benchmark systems and a coculture bioreactor subject to external forcing, as well as for uncertainty quantification of a microbial electrosynthesis reactor. Such data-driven models and analyses of dynamical systems can be paramount in the design and optimization of bioprocesses and integrated biomanufacturing systems.     

Modeling of Hysteresis in Gene Regulatory Networks

Hysteresis, observed in many gene regulatory networks, has a pivotal impact on biological systems, which enhances the robustness of cell functions. In this paper, a general model is proposed to describe the hysteretic gene regulatory network by combining the hysteresis component and the transient dynamics. The Bouc-Wen hysteresis model is modified to describe the hysteresis component in the mammalian gene regulatory networks. Rigorous mathematical analysis on the dynamical properties of the model is presented to ensure the bounded-input-bounded-output (BIBO) stability and demonstrates that the original Bouc-Wen model can only generate a clockwise hysteresis loop while the modified model can describe both clockwise and counter clockwise hysteresis loops. Simulation studies have shown that the hysteresis loops from our model are consistent with the experimental observations in three mammalian gene regulatory networks and two E.coli gene regulatory networks, which demonstrate the ability and accuracy of the mathematical model to emulate natural gene expression behavior with hysteresis. A comparison study has also been conducted to show that this model fits the experiment data significantly better than previous ones in the literature. The successful modeling of the hysteresis in all the five hysteretic gene regulatory networks suggests that the new model has the potential to be a unified framework for modeling hysteresis in gene regulatory networks and provide better understanding of the general mechanism that drives the hysteretic function.     

Bringing sequences to life: how bioinformatics corporealizes sequence data

Sequence databases have become more visible through the heavy publicity associated with the Human Genome Project. This paper looks at some of the emerging artefacts and systems developed to read, write, order and visualise sequence data and other kinds of biological data retrieved from databases and databanks such as GenBank. It argues that bioinformatics software can be regarded as a symptom of a broad and powerful transformation in biological knowledges and in the biopolitical constitution of living bodies. Two facets of this transformation are emphasised. Firstly, examining bioinformatics software might help us situate how sequence data is actually circulating, and to what ends. In particular, the paper looks at the significance of sequence comparison and protein folding problems. Secondly, an emerging nexus of property relations and intellectual work can be detected within the ordering of sequence data carried out bioinformatics.     

A computational modeling and analysis in cell biological dynamics using electric cell-substrate impedance sensing (ECIS)

In this paper, a study of computational modeling and multi-scale analysis in cell dynamics is presented. Our study aims at: (1) deriving and validating a mathematical model for cell growth, and (2) quantitatively detecting and analyzing the biological interdependencies across multiple observational scales with a variety of time and frequency resolutions. This research was conducted using the time series data practically measured from a novel on-line cell monitoring technique, referred to as electric cell-substrate impedance sensing (ECIS), which allows continuously tracking the cellular behavior such as adhesion, proliferation, spreading and micromotion. First, comparing our ECIS-based cellular growth modeling analysis results with those determined by hematocytometer measurement using different time intervals, we found that the results obtained from both experimental methods consistently agreed. However, our study demonstrated that it is much easier and more convenient to operate with the ECIS system for on-line cellular growth monitoring. Secondly, for multi-scale analysis our results showed that the proposed wavelet-based methodology can effectively quantify the fluctuations associated with cell micromotions and quantitatively capture the biological interdependencies across multiple observational scales. Note that although the wavelet method is well known, its application into the ECIS time series analysis is novel and unprecedented in computational cell biology. Our analyses indicated that the proposed study on ECIS time series could provide a hopeful start and great potentials in both modeling and elucidating the complex mechanisms of cell biological systems.     

Pathway analysis using random forests classification and regression

MOTIVATION: Although numerous methods have been developed to better capture biological information from microarray data, commonly used single gene-based methods neglect interactions among genes and leave room for other novel approaches. For example, most classification and regression methods for microarray data are based on the whole set of genes and have not made use of pathway information. Pathway-based analysis in microarray studies may lead to more informative and relevant knowledge for biological researchers. RESULTS: In this paper, we describe a pathway-based classification and regression method using Random Forests to analyze gene expression data. The proposed methods allow researchers to rank important pathways from externally available databases, discover important genes, find pathway-based outlying cases and make full use of a continuous outcome variable in the regression setting. We also compared Random Forests with other machine learning methods using several datasets and found that Random Forests classification error rates were either the lowest or the second-lowest. By combining pathway information and novel statistical methods, this procedure represents a promising computational strategy in dissecting pathways and can provide biological insight into the study of microarray data. AVAILABILITY: Source code written in R is available from http://bioinformatics.med.yale.edu/pathway-analysis/rf.htm.     

Simultaneous Determination of Protein Structure and Dynamics Using Cryo-Electron Microscopy

Cryo-electron microscopy is rapidly emerging as a powerful technique to determine the structures of complex macromolecular systems elusive to other techniques. Because many of these systems are highly dynamical, characterizing their movements is also a crucial step to unravel their biological functions. To achieve this goal, we report an integrative modeling approach to simultaneously determine structure and dynamics of macromolecular systems from cryo-electron microscopy density maps. By quantifying the level of noise in the data and dealing with their ensemble-averaged nature, this approach enables the integration of multiple sources of information to model ensembles of structures and infer their populations. We illustrate the method by characterizing structure and dynamics of the integral membrane receptor STRA6, thus providing insights into the mechanisms by which it interacts with retinol binding protein and translocates retinol across the membrane.     

Dynamics of Boolean networks controlled by biologically meaningful functions

The remarkably stable dynamics displayed by randomly constructed Boolean networks is one of the most striking examples of the spontaneous emergence of self-organization in model systems composed of many interacting elements (Kauffman, S., J. theor. Biol.22, 437-467, 1969; The Origins of Order, Oxford University Press, Oxford, 1993). The dynamics of such networks is most stable for a connectivity of two inputs per element, and decreases dramatically with increasing number of connections. Whereas the simplicity of this model system allows the tracing of the dynamical trajectories, it leaves out many features of real biological connections. For instance, the dynamics has been studied in detail only for networks constructed by allowing all theoretically possible Boolean rules, whereas only a subset of them make sense in the material world. This paper analyses the effect on the dynamics of using only Boolean functions which are meaningful in a biological sense. This analysis is particularly relevant for nets with more than two inputs per element because biological networks generally appear to be more extensively interconnected. Sets of the meaningful functions were assembled for up to four inputs per element. The use of these rules results in a smaller number of distinct attractors which have a shorter length, with relatively little sensitivity to the size of the network and to the number of inputs per element. Forcing away the activator/inhibitor ratio from the expected value of 50% further enhances the stability. This effect is more pronounced for networks consisting of a majority of activators than for networks with a corresponding majority of inhibitors, indicating that the former allow the evolution of larger genetic networks. The data further support the idea of the usefulness of logical networks as a conceptual framework for the understanding of real-world phenomena.     

Empirical orthogonal function regression: Linking population biology to spatial varying environmental conditions using climate projections

Ecologists and oceanographers inform population and ecosystem management by identifying the physical drivers of ecological dynamics. However, different research communities use different analytical tools where, for example, physical oceanographers often apply rank‐reduction techniques (a.k.a. empirical orthogonal functions [EOF]) to identify indicators that represent dominant modes of physical variability, whereas population ecologists use dynamical models that incorporate physical indicators as covariates. Simultaneously modeling physical and biological processes would have several benefits, including improved communication across sub‐fields; more efficient use of limited data; and the ability to compare importance of physical and biological drivers for population dynamics. Here, we develop a new statistical technique, EOF regression, which jointly models population‐scale dynamics and spatially distributed physical dynamics. EOF regression is fitted using maximum‐likelihood techniques and applies a generalized EOF analysis to environmental measurements, estimates one or more time series representing modes of environmental variability, and simultaneously estimates the association of this time series with biological measurements. By doing so, it identifies a spatial map of environmental conditions that are best correlated with annual variability in the biological process. We demonstrate this method using a linear (Ricker) model for early‐life survival (“recruitment”) of three groundfish species in the eastern Bering Sea from 1982 to 2016, combined with measurements and end‐of‐century projections for bottom and sea surface temperature. Results suggest that (a) we can forecast biological dynamics while applying delta‐correction and statistical downscaling to calibrate measurements and projected physical variables, (b) physical drivers are statistically significant for Pacific cod and walleye pollock recruitment, (c) separately analyzing physical and biological variables fails to identify the significant association for walleye pollock, and (d) cod and pollock will likely have reduced recruitment given forecasted temperatures over future decades.