On the Use of Topological Features of Metabolic Networks for the Classification of Cancer Samples.

BACKGROUND: The increasing availability of omics data collected from patients affected by severe pathologies, such as cancer, is fostering the development of data science methods for their analysis. INTRODUCTION: The combination of data integration and machine learning approaches can provide new powerful instruments to tackle the complexity of cancer development and deliver effective diagnostic and prognostic strategies. METHODS: We explore the possibility of exploiting the topological properties of sample-specific metabolic networks as features in a supervised classification task. Such networks are obtained by projecting transcriptomic data from RNA-seq experiments on genome-wide metabolic models to define weighted networks modeling the overall metabolic activity of a given sample. RESULTS: We show the classification results on a labeled breast cancer dataset from the TCGA database, including 210 samples (cancer vs. normal). In particular, we investigate how the performance is affected by a threshold-based pruning of the networks by comparing Artificial Neural Networks, Support Vector Machines and Random Forests. Interestingly, the best classification performance is achieved within a small threshold range for all methods, suggesting that it might represent an effective choice to recover useful information while filtering out noise from data. Overall, the best accuracy is achieved with SVMs, which exhibit performances similar to those obtained when gene expression profiles are used as features. CONCLUSION: These findings demonstrate that the topological properties of sample-specific metabolic networks are effective in classifying cancer and normal samples, suggesting that useful information can be extracted from a relatively limited number of features.

Topological assessment of metabolic networks reveals evolutionary information.

Evolutionary information was inferred from the topology of metabolic networks corresponding to 17 plant species belonging to major plant lineages Chlorophytes, Bryophytes, Lycophytes and Angiosperms. The plant metabolic networks were built using the substrate-product network modeling based on the metabolic reactions available on the PlantCyc database (version 9.5), from which their local topological properties such as degree, in-degree, out-degree, clustering coefficient, hub-score, authority-score, local efficiency, betweenness and eigencentrality were measured. The topological measurements corresponding to each metabolite within the networks were considered as a set of metabolic characters to compound a feature vector representing each plant. Our results revealed that some local topological characters are able to discern among plant kinships, since similar phylogenies were found when comparing dendrograms obtained by topological metrics to the one obtained by DNA sequences of chloroplast genes. Furthermore, we also found that even a smaller number of metabolic characters is able to separate among major clades with high bootstrap support (BS > 95), while for some suborders a bigger content has been required.

A hierarchical model of metabolic machinery based on the kcore decomposition of plant metabolic networks.

Modeling the basic structure of metabolic machinery is a challenge for modern biology. Some models based on complex networks have provided important information regarding this machinery. In this paper, we constructed metabolic networks of 17 plants covering unicellular organisms to more complex dicotyledonous plants. The metabolic networks were built based on the substrate-product model and a topological percolation was performed using the kcore decomposition. The distribution of metabolites across the percolation layers showed correlations between the metabolic integration hierarchy and the network topology. We show that metabolites concentrated in the internal network (maximum kcore) only comprise molecules of the primary basal metabolism. Moreover, we found a high proportion of a set of common metabolites, among the 17 plants, centered at the inner kcore layers. Meanwhile, the metabolites recognized as participants in the secondary metabolism of plants are concentrated in the outermost layers of the network. This data suggests that the metabolites in the central layer form a basic molecular module in which the whole plant metabolism is anchored. The elements from this central core participate in almost all plant metabolic reactions, which suggests that plant metabolic networks follows a centralized topology.

Authorship attribution based on Life-Like Network Automata.

The authorship attribution is a problem of considerable practical and technical interest. Several methods have been designed to infer the authorship of disputed documents in multiple contexts. While traditional statistical methods based solely on word counts and related measurements have provided a simple, yet effective solution in particular cases; they are prone to manipulation. Recently, texts have been successfully modeled as networks, where words are represented by nodes linked according to textual similarity measurements. Such models are useful to identify informative topological patterns for the authorship recognition task. However, there is no consensus on which measurements should be used. Thus, we proposed a novel method to characterize text networks, by considering both topological and dynamical aspects of networks. Using concepts and methods from cellular automata theory, we devised a strategy to grasp informative spatio-temporal patterns from this model. Our experiments revealed an outperformance over structural analysis relying only on topological measurements, such as clustering coefficient, betweenness and shortest paths. The optimized results obtained here pave the way for a better characterization of textual networks.

Improving the pseudo-randomness properties of chaotic maps using deep-zoom.

A generalized method is proposed to compose new orbits from a given chaotic map. The method provides an approach to examine discrete-time chaotic maps in a "deep-zoom" manner by using k-digits to the right from the decimal separator of a given point from the underlying chaotic map. Interesting phenomena have been identified. Rapid randomization was observed, i.e., chaotic patterns tend to become indistinguishable when compared to the original orbits of the underlying chaotic map. Our results were presented using different graphical analyses (i.e., time-evolution, bifurcation diagram, Lyapunov exponent, Poincaré diagram, and frequency distribution). Moreover, taking advantage of this randomization improvement, we propose a Pseudo-Random Number Generator (PRNG) based on the k-logistic map. The pseudo-random qualities of the proposed PRNG passed both tests successfully, i.e., DIEHARD and NIST, and were comparable with other traditional PRNGs such as the Mersenne Twister. The results suggest that simple maps such as the logistic map can be considered as good PRNG methods.

Exploring Spatio-temporal Dynamics of Cellular Automata for Pattern Recognition in Networks.

Network science is an interdisciplinary field which provides an integrative approach for the study of complex systems. In recent years, network modeling has been used for the study of emergent phenomena in many real-world applications. Pattern recognition in networks has been drawing attention to the importance of network characterization, which may lead to understanding the topological properties that are related to the network model. In this paper, the Life-Like Network Automata (LLNA) method is introduced, which was designed for pattern recognition in networks. LLNA uses the network topology as a tessellation of Cellular Automata (CA), whose dynamics produces a spatio-temporal pattern used to extract the feature vector for network characterization. The method was evaluated using synthetic and real-world networks. In the latter, three pattern recognition applications were used: (i) identifying organisms from distinct domains of life through their metabolic networks, (ii) identifying online social networks and (iii) classifying stomata distribution patterns varying according to different lighting conditions. LLNA was compared to structural measurements and surpasses them in real-world applications, achieving improvement in the classification rate as high as 23%, 4% and 7% respectively. Therefore, the proposed method is a good choice for pattern recognition applications using networks and demonstrates potential for general applicability.