Article
Chemistry, Medicinal
Zhenqiu Shu, Qinghan Long, Luping Zhang, Zhengtao Yu, Xiao-Jun Wu
Summary: The progress in single-cell RNA sequencing (ScRNA-seq) technology allows for the accurate discovery of cell heterogeneity and diversity. Clustering is a crucial step in ScRNA-seq data analysis, but it faces challenges due to the high dimensionality and noise of the data. To overcome these challenges, we propose a novel ScRNA-seq data clustering model, RGNMF-DS, which incorporates similarity and dissimilarity regularizers for matrix decomposition and utilizes a graph regularizer to uncover the local geometric structure in the data. Experimental results demonstrate that our proposed model outperforms other state-of-the-art methods in clustering ScRNA-seq datasets.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2022)
Review
Biochemical Research Methods
Junkang Wei, Siyuan Chen, Licheng Zong, Xin Gao, Yu Li
Summary: Protein-RNA interactions play a vital role in cellular activities. Previous computational methods heavily rely on sequence data due to the lack of protein structure data. However, the emergence of AlphaFold is set to revolutionize protein-RNA interaction prediction. In this review, we provide a comprehensive overview of the field, covering binding site and binding preference prediction, as well as commonly used datasets, features, and models. We also discuss potential challenges and opportunities in this area.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Engineering, Environmental
Mainak Chatterjee, Kunal Roy
Summary: This paper developed QSAR models for predicting aquatic toxicity, using Partial Least Squares regression as a statistical tool. The models were based on structural features of individual chemicals and mixture components, with quality assessed by strict validation parameters. The final models are robust, highly predictive, and mechanistically interpretable for predicting toxicity of untested chemical mixtures within the domain of applicability.
JOURNAL OF HAZARDOUS MATERIALS
(2021)
Article
Genetics & Heredity
Osval A. Montesinos-Lopez, Abelardo Montesinos-Lopez, David Alejandro Bernal Sandoval, Brandon Alejandro Mosqueda-Gonzalez, Marco Alberto Valenzo-Jimenez, Jose Crossa
Summary: The genomic selection methodology has revolutionized plant breeding by using statistical machine learning algorithms to predict candidate individuals. However, it faces challenges when predicting future seasons or new environments. This study compared the performance of the multi-trait partial least square (MT-PLS) regression method with the Bayesian Multi-trait Genomic Best Linear Unbiased Predictor (MT-GBLUP) method and found that MT-PLS outperforms MT-GBLUP in predicting future seasons or new environments.
FRONTIERS IN GENETICS
(2022)
Article
Chemistry, Analytical
An-Qi He, Zhen-Qiang Yu, Jun Song, Li-Min Yang, Yi-Zhuang Xu, Isao Noda, Yukihiro Ozaki
Summary: A novel spectral analysis approach is proposed to retrieve the spectrum of a supramolecular complex. The approach involves constructing a two-dimensional asynchronous spectrum and using a genetic algorithm to obtain a heuristic spectrum of the complex. The results suggest that this approach provides a new insight into studying various intermolecular interactions.
ANALYTICAL CHEMISTRY
(2022)
Article
Chemistry, Multidisciplinary
Sai Zou, Yunbin Hu, Wenya Yang
Summary: Identifying essential proteins is crucial for understanding cellular requirements, discovering pathogenic genes, and diagnosing diseases. The integration of protein-protein interaction networks and biological sequence features enhances the accuracy of essential protein identification. A deep neural network method named IYEPDNN was used in this study, achieving a high accuracy of 84% and outperforming other state-of-the-art methods.
APPLIED SCIENCES-BASEL
(2023)
Article
Biochemical Research Methods
Zhourun Wu, Qing Liao, Shixi Fan, Bin Liu
Summary: A novel computational method idenPC-CAP was developed to identify protein complexes from the RNA-protein heterogeneous interaction network, reducing the false positive proteins ratio and outperforming other state-of-the-art methods in this field as demonstrated by experimental results.
BRIEFINGS IN BIOINFORMATICS
(2021)
Article
Multidisciplinary Sciences
Md. Nahid Pervez, Wan Sieng Yeo, Mst. Monira Rahman Mishu, Md. Eman Talukder, Hridoy Roy, Md. Shahinoor Islam, Yaping Zhao, Yingjie Cai, George K. Stylios, Vincenzo Naddeo
Summary: Despite limited simulation studies, this research developed a sustainable and effective electrospinning process by combining the design of experiments with machine learning prediction models. The locally weighted kernel partial least squares regression (LW-KPLSR) model, based on response surface methodology, outperformed other models in predicting the membrane diameter.
SCIENTIFIC REPORTS
(2023)
Article
Food Science & Technology
Jiayi Hang, Da Shi, Jason Neufeld, Kirstin E. Bett, James D. House
Summary: This study developed near-infrared reflectance spectroscopy models to predict the protein and amino acid contents in lentil seeds. The results showed that the models achieved good statistical results and had the potential for rapid and accurate prediction of these nutritional components in lentils.
LWT-FOOD SCIENCE AND TECHNOLOGY
(2022)
Article
Multidisciplinary Sciences
Qinqing Xiong, Wenju Wang, Mingya Wang, Chunhui Zhang, Xuechun Zhang, Chun Chen, Mingshi Wang
Summary: This study proposes a hybrid neural network model SOM-NARX based on the correlation of predictors for ozone prediction. The model filters predictors using MIC, transforms them into feature sequences using SOM, and makes predictions using NARX networks. The results show that the correlation of predictors, classification numbers of SOM, neuron numbers, and delay steps can affect prediction accuracy. Model comparison shows that the SOM-NARX model outperforms other models in terms of RMSE, MAE, and MAEP.
Article
Multidisciplinary Sciences
Xiangyu Guo, Ahmed Jahoor, Just Jensen, Pernille Sarup
Summary: In this study, metabolomic spectra were used to predict malting quality phenotypes in different locations, and the prediction ability of different models and training population sizes were compared. The results showed that more than 90% of the total variance in malting quality traits could be explained by metabolomic features. The prediction accuracy increased with increasing training population size and stabilized when the size reached 1000. The optimal number of components considered in the prediction models was 20. The accuracy using cross-validation ranged from 0.722 to 0.865 for leave-one-line-out and from 0.517 to 0.817 for leave-one-location-out. Therefore, metabolomic prediction of malting quality traits using metabolomic features has high accuracy, and MBLUP is better than PLSR when the training population size is larger than 100.
SCIENTIFIC REPORTS
(2022)
Article
Biology
Xiujuan Zhao, Yanping Zhang, Xiuquan Du
Summary: DFpin is a method for predicting protein-interacting nucleotides in RNA. It removes redundancy based on feature similarity and uses a deep forest model to extract key features, achieving an accuracy of 85.4%.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Mathematics, Applied
Bjorn Engquist, Yunan Yang
Summary: Full-waveform inversion is a standard process for seismic imaging inverse problems, using PDE-constrained optimization to determine unknown geophysical parameters. Introducing the Wasserstein metric can help mitigate cycle skipping and improve inversion accuracy.
COMMUNICATIONS ON PURE AND APPLIED MATHEMATICS
(2022)
Article
Chemistry, Medicinal
Francini Fonseca Lopez, Jiayuan Miao, Jovan Damjanovic, Luca Bischof, Michael B. Braun, Yingjie Ling, Marcus D. Hartmann, Yu-Shan Lin, Joshua A. Kritzer
Summary: The Nrf2 transcription factor regulates the response to oxidative stress, and Keap1 is its main negative regulator. This study used molecular dynamics simulations to predict the preorganization of cyclic peptides and correlated it with binding affinities for Keap1. The findings provide insights into designing selective inhibitors of protein-protein interactions using cyclic peptides.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2023)
Article
Thermodynamics
Wente Niu, Jialiang Lu, Yuping Sun, Wei Guo, Yuyang Liu, Ying Mu
Summary: This study clarifies the main controlling factors of EUR for shale gas wells through sensitivity analysis. Visual forecasting models were established using different methods and the model based on LSSVM showed the best field application effect. The proposed model is a reliable and efficient tool for EUR prediction.