In high-dimensional data, the performances of various classifiers are largely dependent on the selection of important features. Most of the individual classifiers with the existing feature selection (FS) methods do not perform well for highly correlated data. Obtaining important features using the FS method and selecting the best performing classifier is a challenging task in high throughput data. In this article, we propose a combination of resampling-based least absolute shrinkage and selection operator (LASSO) feature selection (RLFS) and ensembles of regularized regression (ERRM) capable of dealing data with the high correlation structures. The ERRM boosts the prediction accuracy with the top-ranked features obtained from RLFS. The RLFS utilizes the lasso penalty with sure independence screening (SIS) condition to select the top k ranked features. The ERRM includes five individual penalty based classifiers: LASSO, adaptive LASSO (ALASSO), elastic net (ENET), smoothly clipped absolute deviations (SCAD), and minimax concave penalty (MCP). It was built on the idea of bagging and rank aggregation. Upon performing simulation studies and applying to smokers’ cancer gene expression data, we demonstrated that the proposed combination of ERRM with RLFS achieved superior performance of accuracy and geometric mean.

Substantial increases in the conjugation of the main human SUMO paralogs, SUMO1, SUMO2, and SUMO3, are observed upon exposure to different cellular stressors, and such increases are considered important to facilitate cell survival to stress. Despite their critical cellular role, little is known about how the levels of the SUMO modifiers are regulated in the cell, particularly as it relates to the changes observed upon stress. Here we characterize the contribution of alternative splicing towards regulating the expression of the main human SUMO paralogs under normalcy and three different stress conditions, heat-shock, cold-shock, and Influenza A Virus infection. Our data reveal that the normally spliced transcript variants are the predominant mature mRNAs produced from the SUMO genes and that the transcript coding for SUMO2 is by far the most abundant of all. We also provide evidence that alternatively spliced transcripts coding for protein isoforms of the prototypical SUMO proteins, which we refer to as the SUMO alphas, are also produced, and that their abundance and nuclear export are affected by stress in a stress- and cell-specific manner. Additionally, we provide evidence that the SUMO alphas are actively synthesized in the cell as their coding mRNAs are found associated with translating ribosomes. Finally, we provide evidence that the SUMO alphas are functionally different from their prototypical counterparts, with SUMO1α and SUMO2α being non-conjugatable to protein targets, SUMO3α being conjugatable but targeting a seemingly different subset of protein from those targeted by SUMO3, and all three SUMO alphas displaying different cellular distributions from those of the prototypical SUMOs. Thus, alternative splicing appears to be an important contributor to the regulation of the expression of the SUMO proteins and the cellular functions of the SUMOylation system.

Single-cell RNA sequencing (scRNA-seq) is an attractive technology for researchers to gain valuable insights into the cellular processes and cell type diversity present in all tissues. The data generated by the scRNA-seq experiment are high-dimensional and complex in nature. Several tools are now available to analyze the raw scRNA-seq data from public databases; however, simple and easy-to-explore single-cell gene expression visualization tools focusing on differential expression and co-expression are lacking. Here, we present scViewer, an interactive graphical user interface (GUI) R/Shiny application designed to facilitate the visualization of scRNA-seq gene expression data. With the processed Seurat RDS object as input, scViewer utilizes several statistical approaches to provide detailed information on the loaded scRNA-seq experiment and generates publication-ready plots. The major functionalities of scViewer include exploring cell-type-specific gene expression, co-expression analysis of two genes, and differential expression analysis with different biological conditions considering both cell-level and subject-level variations using negative binomial mixed modeling. We utilized a publicly available dataset (brain cells from a study of Alzheimer’s disease to demonstrate the utility of our tool. scViewer can be downloaded from GitHub as a Shiny app with local installation. Overall, scViewer is a user-friendly application that will allow researchers to visualize and interpret the scRNA-seq data efficiently for multi-condition comparison by performing gene-level differential expression and co-expression analysis on the fly. Considering the functionalities of this Shiny app, scViewer can be a great resource for collaboration between bioinformaticians and wet lab scientists for faster data visualizations.

The Illumina Infinium HumanMethylation27 (Illumina 27K) BeadChip assay is a relatively recent high-throughput technology that allows over 27,000 CpGs to be assayed. The Illumina 27K methylation data is less commonly used in comparison to gene expression in bioinformatics. It provides a critical need to find the optimal feature ranking (FR) method for handling the high dimensional data. The optimal FR method on the classifier is not well known, and choosing the best performing FR method becomes more challenging in high dimensional data setting. Therefore, identifying the statistical methods which boost the inference is of crucial importance in this context. This paper describes the detailed performances of FR methods such as fisher score, information gain, chi-square, and minimum redundancy and maximum relevance on different classification methods such as Adaboost, Random Forest, Naive Bayes, and Support Vector Machines. Through simulation study and real data applications, we show that the fisher score as an FR method, when applied on all the classifiers, achieved best prediction accuracy with significantly small number of ranked features.