FeatureCorr: An R package to study feature correlations aided with data transformation for sequencing and microarray data

Francesca Storici,Deepali Kundnani

doi:10.1016/j.simpa.2021.100144

Abstract

<h2>Abstract</h2> FeatureCorr is an R package that aids in association and network analysis of data obtained from preliminary bioinformatic analysis of next-generation sequencing(NGS) or microarray experiments. These experiments are widely used for various applications like mutation and expression profiles, detection of epigenetic changes in the genomic DNA, etc. FeatureCorr enables users in cleaning and preprocessing data to minimize batch effects and for background noise removal. FeatureCorr can help in the analysis of feature correlation in different ways: Correlation of one Feature vs multiple Features, pairwise correlation of multiple features against multiple features, and in-depth correlation and distributions of two features.

Highlights

Next-Generation Sequencing (NGS) and microarray experiments have evolved rapidly after the completion of The Human GenomeProject and novel methods are continually being commercialized [1]
FeatureCorr is an R package that works with a two dimensional R Data frame containing quantified values or signals from preliminary bioinformatic analysis of NGS experiments to generate associations between features using Spearman correlation for non-parametric analysis and Pearson correlation which uses linear regression for parametric correlation analysis
Different transformations for Sequencing data with distribution visualization of original and transformed data can aid in more accurate correlations between features of interest

Summary

Introduction

Next-Generation Sequencing (NGS) and microarray experiments have evolved rapidly after the completion of The Human Genome. The code (and data) in this article has been certified as Reproducible by Code Ocean: (https://codeocean.com/). Storici by collecting Sequencing and microarray data from different research groups and made use of automated bioinformatic pipelines to generate preliminary quantification data for NGS experiments like DNA and RNA sequencing [2]. The preliminary data is quantified for features (identifiers) associated with genomic locations or genomic properties such as genes, transcription factor binding sites, etc. FeatureCorr is an R package that works with a two dimensional R Data frame containing quantified values or signals from preliminary bioinformatic analysis of NGS experiments to generate associations between features using Spearman correlation for non-parametric analysis and Pearson correlation which uses linear regression for parametric correlation analysis. The user can visualize distribution between specific pairs of features using FeatureCorr for deeper understanding

FeatureCorr Package Functions

Correlation Analysis

Conclusion

Impact