Article
Biology
Claude Pasquier, Vincent Guerlais, Denis Pallez, Raphael Rapetti-Mauss, Olivier Soriani
Summary: The identification of condition-specific gene sets is crucial in understanding regulatory and signaling mechanisms associated with cellular response. Existing differential expression analysis methods have limitations in identifying small varying gene modules that play significant roles in phenotypic changes. In this study, we propose an efficient method that combines gene expressions and interaction data to identify these active modules. Our method has been shown to uncover new groups of functionally important genes that are not captured by traditional approaches. The software is available at https://github.com/claudepasquier/amine.
LIFE SCIENCE ALLIANCE
(2023)
Article
Computer Science, Artificial Intelligence
Houssem Eddine Nouri, Abdennaceur Ghandri, Olfa Belkahla Driss, Khaled Ghedira
Summary: This paper proposes a new design of evolutionary algorithms, called Bi-level Evolutionary Approach (Bi-EvoGAN), to solve the optimization problems of generative adversarial networks (GANs). By dividing the evolution search space into two complementary levels, Bi-EvoGAN is able to optimize both the topology and hyperparameters of GANs, thereby improving the quality of the search method and accelerating convergence. Experimental results demonstrate the superiority of Bi-EvoGAN over other state-of-the-art approaches on four benchmark datasets.
APPLIED SOFT COMPUTING
(2023)
Article
Multidisciplinary Sciences
Gianalberto Losapio, Christian Schoeb, Phillip P. A. Staniczenko, Francesco Carrara, Gian Marco Palamara, Consuelo M. De Moraes, Mark C. Mescher, Rob W. Brooker, Bradley J. Butterfield, Ragan M. Callaway, Lohengrin A. Cavieres, Zaal Kikvidze, Christopher J. Lortie, Richard Michalet, Francisco Pugnaire, Jordi Bascompte
Summary: The study found that patterns of positive and negative associations among species in alpine plant populations have a positive impact on species diversity globally, contributing to the persistence of local communities. This highlights the importance of competition and facilitation in maintaining biodiversity.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2021)
Article
Computer Science, Information Systems
Muhammad Abrar Afzal, Zhenyu Gu, Bilal Afzal, Syed Umer Bukhari
Summary: In the era of Industry 5.0, effectively managing cognitive workload is crucial for optimizing human performance and ensuring operational efficiency. This study proposes an EEG-based Bi-directional Gated Network (BDGN) approach that incorporates LSTM and GRU models for cognitive workload classification in Industry 5.0 applications. The research demonstrates an impressive accuracy of 98% in classifying cognitive workload using the suggested BDGN approach.
Article
Environmental Sciences
Seong-Jun Chun, Young-Joong Kim, Yingshun Cui, Kyong-Hee Nam
Summary: Heavy metal pollution in soil around abandoned mine sites is a critical environmental issue worldwide. In this study, the distribution patterns of bacterial and fungal communities in non-contaminated and heavy metal-contaminated soil were investigated to explore microbial interaction mechanisms and their modular structures. The microbial network was divided into three modules based on the levels of heavy metal pollution, with copper playing a key role in the formation of the heavy metal-tolerant module. Overall, distinct microbial communities formed in response to heavy metal contamination, potentially contributing to bioremediation efforts.
ENVIRONMENTAL POLLUTION
(2021)
Article
Biochemical Research Methods
Thao Vu, Elizabeth M. Litkowski, Weixuan Liu, Katherine A. Pratte, Leslie Lange, Russell P. Bowler, Farnoush Banaei-Kashani, Katerina J. Kechris
Summary: Biological networks provide a system-level understanding of underlying processes. In order to investigate the association between modules and variables of interest, a module summarization method that can explain the module's information and reduce dimensionality is needed. This article proposes NetSHy, a hybrid approach that incorporates topological properties to aid downstream analysis interpretation while reducing network dimensionality.
Article
Multidisciplinary Sciences
Priyanka Kumari, Bikram Pradhan, Maria Koromina, George P. Patrinos, Kristel Van Steen
Summary: The outbreak of COVID-19 has created a global threat to public health, and there is a need to find effective and safe treatments. Dexamethasone, baricitinib, and remdesivir have shown high antiviral activity against SARS-CoV-2 in vitro. This study aims to explore the potential uses of these drugs through in-silico research.
Article
Automation & Control Systems
Qie Liu, Biao Huang, Yi Chai, Wenbo Li
Summary: The identification of topology in sparse networks is crucial for network modeling in various fields. An efficient algorithm based on stochastic optimization is proposed to decrease computational complexity and is suitable for network identification with large data sets.
Article
Biochemical Research Methods
Yujie Wang, Gang Zhou, Tianhao Guan, Yan Wang, Chenxu Xuan, Tao Ding, Jie Gao
Summary: This study proposes a network-regularized sparse orthogonal-regularized joint non-negative matrix factorization (NSOJNMF) algorithm to explore RNA expression patterns and potential molecular mechanisms of cancer. By applying this algorithm, ceRNA co-modules of liver cancer and colon cancer are identified, and enrichment analysis shows their close association with cancer occurrence and development. In addition, potential biological associations among RNA molecules are discovered through the constructed ceRNA networks.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Chemistry, Analytical
Bo Peng, Yuanming Ding, Wei Kang
Summary: Since the introduction of the Transformer model, it has had a significant impact on various fields of machine learning, including time series prediction. However, the existing multi-head attention mechanisms suffer from issues of feature redundancy and resource waste. To address these problems, this paper proposes a hierarchical attention mechanism and utilizes global feature aggregation using graph networks to enhance the information diversity and improve performance.
Article
Mathematics, Interdisciplinary Applications
Daniel Straulino, Mattie Landman, Neave O'Clery
Summary: This study proposes a new method to compare the modular structure of a pair of node-aligned networks by assessing the fit of each node partition with respect to the other network's connectivity structure. The method is adaptable to various community detection algorithms, takes into account network structure, and can identify differences in networks with similar partitions but varying community structures.
Article
Engineering, Industrial
M. A. S. Monfared, Masoumeh Rezazadeh, Zohreh Alipour
Summary: This paper presents a novel three-stage heuristic algorithm to evaluate and improve the reliability of large-scale road networks. The method is based on mathematical theory and utilizes bottom-up and top-down approaches to enhance network reliability. The research demonstrates the efficiency and feasibility of the proposed method.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2022)
Review
Biochemistry & Molecular Biology
Ugo Dionne, Lily J. Percival, Francois J. M. Chartier, Christian R. Landry, Nicolas Bisson
Summary: The assembly of complexes following the detection of extracellular signals is controlled by signaling proteins with multiple peptide binding modules. The SH3 family, with around 300 annotated SH3 domains in humans, is a modular protein interaction module that regulates various signaling processes. Recent findings have challenged the simple model of SH3 domains as portable domains binding to specific proline-rich peptide motifs, by revealing their allosteric contributions in the host protein context, phosphoregulation, and roles in phase separation.
TRENDS IN BIOCHEMICAL SCIENCES
(2022)
Article
Engineering, Geological
Josephine Morgenroth, Matthew A. Perras, Usman T. Khan
Summary: Advancements in underground instrumentation and cost-effective data storage have provided the rock engineering community with larger datasets than ever before. Machine learning algorithms, specifically Convolutional Neural Networks, offer a more efficient way to analyze rock mass deformation mechanics, ultimately increasing the reliability of underground excavations.
ROCK MECHANICS AND ROCK ENGINEERING
(2022)
Article
Engineering, Multidisciplinary
S. Naveen Venkatesh, V. Sugumaran
Summary: This study aims to identify visual faults in photovoltaic modules using machine vision and machine learning techniques, specifically through the classification of normal RGB images with the fusion of deep learning and machine learning methods.