Inferring a gene regulatory network from time-series gene expression data in systems biology is a challenging problem. Many methods have been suggested, most of which have a scalability limitation due to the combinatorial cost of searching a regulatory set of genes. In addition, they have focused on the accurate inference of a network structure only. Therefore, there is a pressing need to develop a network inference method to search regulatory genes efficiently and to predict the network dynamics accurately. In this study, we employed a Boolean network model with a restricted update rule scheme to capture coarse-grained dynamics, and propose a novel mutual information-based Boolean network inference (MIBNI) method. Given time-series gene expression data as an input, the method first identifies a set of initial regulatory genes using mutual information-based feature selection, and then improves the dynamics prediction accuracy by iteratively swapping a pair of genes between sets of the selected regulatory genes and the other genes. Through extensive simulations with artificial datasets, MIBNI showed consistently better performance than six well-known existing methods, REVEAL, Best-Fit, RelNet, CST, CLR, and BIBN in terms of both structural and dynamics prediction accuracy. We further tested the proposed method with two real gene expression datasets for an Escherichia coli gene regulatory network and a fission yeast cell cycle network, and also observed better results using MIBNI compared to the six other methods. Taken together, MIBNI is a promising tool for predicting both the structure and the dynamics of a gene regulatory network.

It has been a challenge in systems biology to unravel relationships between structural properties and dynamic behaviors of biological networks. A Cytoscape plugin named NetDS was recently proposed to analyze the robustness-related dynamics and feed-forward/feedback loop structures of biological networks. Despite such a useful function, limitations on the network size that can be analyzed exist due to high computational costs. In addition, the plugin cannot verify an intrinsic property which can be induced by an observed result because it has no function to simulate the observation on a large number of random networks. To overcome these limitations, we have developed a novel software tool, PANET. First, the time-consuming parts of NetDS were redesigned to be processed in parallel using the OpenCL library. This approach utilizes the full computing power of multi-core central processing units and graphics processing units. Eventually, this made it possible to investigate a large-scale network such as a human signaling network with 1,609 nodes and 5,063 links. We also developed a new function to perform a batch-mode simulation where it generates a lot of random networks and conducts robustness calculations and feed-forward/feedback loop examinations of them. This helps us to determine if the findings in real biological networks are valid in arbitrary random networks or not. We tested our plugin in two case studies based on two large-scale signaling networks and found interesting results regarding relationships between coherently coupled feed-forward/feedback loops and robustness. In addition, we verified whether or not those findings are consistently conserved in random networks through batch-mode simulations. Taken together, our plugin is expected to effectively investigate various relationships between dynamics and structural properties in large-scale networks. Our software tool, user manual and example datasets are freely available at http://panet-csc.sourceforge.net/.

Many biological networks tend to have a high modularity structural property and the dynamic characteristic of high robustness against perturbations. However, the relationship between modularity and robustness is not well understood. To investigate this relationship, we examined real signalling networks and conducted simulations using a random Boolean network model. As a result, we first observed that the network robustness is negatively correlated with the network modularity. In particular, this negative correlation becomes more apparent as the network density becomes sparser. Even more interesting is that, the negative relationship between the network robustness and the network modularity occurs mainly because nodes in the same module with the perturbed node tend to be more sensitive to the perturbation than those in other modules. This result implies that dynamically similar nodes tend to be located in the same module of a network. To support this, we show that a pair of genes associated with the same disease or a pair of functionally similar genes is likely to belong to the same module in a human signalling network.

It is a challenging problem in systems biology to infer both the network structure and dynamics of a gene regulatory network from steady-state gene expression data. Some methods based on Boolean or differential equation models have been proposed but they were not efficient in inference of large-scale networks. Therefore, it is necessary to develop a method to infer the network structure and dynamics accurately on large-scale networks using steady-state expression. In this study, we propose a novel constrained genetic algorithm-based Boolean network inference (CGA-BNI) method where a Boolean canalyzing update rule scheme was employed to capture coarse-grained dynamics. Given steady-state gene expression data as an input, CGA-BNI identifies a set of path consistency-based constraints by comparing the gene expression level between the wild-type and the mutant experiments. It then searches Boolean networks which satisfy the constraints and induce attractors most similar to steady-state expressions. We devised a heuristic mutation operation for faster convergence and implemented a parallel evaluation routine for execution time reduction. Through extensive simulations on the artificial and the real gene expression datasets, CGA-BNI showed better performance than four other existing methods in terms of both structural and dynamics prediction accuracies. Taken together, CGA-BNI is a promising tool to predict both the structure and the dynamics of a gene regulatory network when a highest accuracy is needed at the cost of sacrificing the execution time. Source code and data are freely available at https://github.com/csclab/CGA-BNI. Supplementary data are available at Bioinformatics online.

There have been many machine learning-based studies to forecast stock price trends. These studies attempted to extract input features mostly from the price information with little focus on the trading volume information. In addition, modeling parameters to specify a learning problem have not been intensively investigated. We herein develop an improved method by handling those limitations. Specifically, we generated input variables by considering both price and volume information with even weight. We also defined three modeling parameters: the input and the target window sizes and the profit threshold. These specify the input and target variables, between which the underlying functions are learned by multilayer perceptrons and support vector machines. We tested our approach over six stocks and 15 years and compared with the expected performance over all considered parameter specifications. Our approach dramatically improved the prediction accuracy over the expected performance. In addition, our approach was shown to be stably more profitable than both the expected performance and the buy-and-hold strategy. On the other hand, the performance was degraded when the input variables generated from the trading volume were excluded from learning. All these results validate the importance of the volume and the modeling parameters in stock trading prediction.

Machinery diagnostics and prognostics usually involve the prediction process of fault-types and remaining useful life (RUL) of a machine, respectively. The process of developing a data-driven diagnostics and prognostics method involves some fundamental subtasks such as data rebalancing, feature extraction, dimension reduction, and machine learning. In general, the best performing algorithm and the optimal hyper-parameters suitable for each subtask are varied across the characteristics of datasets. Therefore, it is challenging to develop a general diagnostic/prognostic framework that can automatically identify the best subtask algorithms and the optimal involved parameters for a given dataset. To resolve this problem, we propose a new framework based on an ensemble of genetic algorithms (GAs) that can be used for both the fault-type classification and RUL prediction. Our GA is combined with a specific machine-learning method and then tries to select the best algorithm and optimize the involved parameter values in each subtask. In addition, our method constructs an ensemble of various prediction models found by the GAs. Our method was compared to a traditional grid-search over three benchmark datasets of the fault-type classification and the RUL prediction problems and showed a significantly better performance than the latter. Taken together, our framework can be an effective approach for the fault-type and RUL prediction of various machinery systems.