A Review of Matched-pairs Feature Selection Methods for Gene Expression Data Analysis

Abstract

With the rapid accumulation of gene expression data from various technologies, e.g., microarray, RNA-sequencing (RNA-seq), and single-cell RNA-seq, it is necessary to carry out dimensional reduction and feature (signature genes) selection in support of making sense out of such high dimensional data. These computational methods significantly facilitate further data analysis and interpretation, such as gene function enrichment analysis, cancer biomarker detection, and drug targeting identification in precision medicine. Although numerous methods have been developed for feature selection in bioinformatics, it is still a challenge to choose the appropriate methods for a specific problem and seek for the most reasonable ranking features. Meanwhile, the paired gene expression data under matched case-control design (MCCD) is becoming increasingly popular, which has often been used in multi-omics integration studies and may increase feature selection efficiency by offsetting similar distributions of confounding features. The appropriate feature selection methods specifically designed for the paired data, which is named as matched-pairs feature selection (MPFS), however, have not been maturely developed in parallel. In this review, we compare the performance of 10 feature-selection methods (eight MPFS methods and two traditional unpaired methods) on two real datasets by applied three classification methods, and analyze the algorithm complexity of these methods through the running of their programs. This review aims to induce and comprehensively present the MPFS in such a way that readers can easily understand its characteristics and get a clue in selecting the appropriate methods for their analyses.

Multi-Stage Feature Selection (MSFS) Algorithm for UWB-Based Early Breast Cancer Size Prediction

“…Data normalization is a method to standardize the range of features without reducing the dimension of the data [19][20][21][22]41]. Data normalization process is important since it is important to select the best features without eliminating useful information from the preprocessed data [19][20][21][22]. Conventional single stage feature selection having the drawback of possibly selecting data after eliminating useful data during feature extraction stage.…”

“…Thus, for this work, raw data samples are normalized using ten different data normalization methods. Based on the comprehensive review done on the previous researches, five data normalization methods are chosen from the commonly used methods, namely, Decimal Scaling (DS), Z-score (ZS), Linear Scaling (LS), Min-Max (MM) and Mean & Standard Deviation (MSD) methods [19][20][21][22]. The other five data normalization methods are newly introduced in early breast cancer detection application, namely, Relative Logarithmic Sum Squared Voltage (RLSSV), Relative Logarithmic Voltage (RLV), Relative Voltage (RV), Fractional Voltage Change (FVC) and Relative Sum Squared Voltage (RSSV) [8][9].…”

“…Once the data is normalized, the normalized data is dimensionally reduced to remove redundant and statistically insignificant data [21]. The dimension of data is reduced as follows:…”

“…Each fold will take turn to be the testing fold, until the training-testing process completed. Confusion matrices are generated for each iteration, and the accuracy, sensitivity and specificity are calculated for each iteration using equations (21) to (23). The average classification accuracy, sensitivity and specificity of all folds are considered as the performance of the classifier [8].…”

Developing a Multi-Layer Deep Learning Based Predictive Model to Identify DNA N4-Methylcytosine Modifications

“…There are currently four methods available in literature to identify 4mC sites, including iDNA4mC (Chen et al, 2017), 4mCPred (Su et al, 2018), 4mcPred-SVM (Wei et al, 2018a), and 4mcPred-IFL (Wei et al, 2019a). iDNA4mC, as the first machine learning predictor, encodes sequences by nucleotide chemical properties and nucleotide frequency to features and trains support vector machine (SVM) models for prediction (Liang et al, 2018). Although this method has the ability to distinguish between 4mC and non-4mC sites, the prediction accuracy is relatively low overall.…”