Removal Of Unwanted Variation Research Articles

SiFT: uncovering hidden biological processes by probabilistic filtering of single-cell data

Cellular populations simultaneously encode multiple biological attributes, including spatial configuration, temporal trajectories, and cell-cell interactions. Some of these signals may be overshadowed by others and harder to recover, despite the great progress made to computationally reconstruct biological processes from single-cell data. To address this, we present SiFT, a kernel-based projection method for filtering biological signals in single-cell data, thus uncovering underlying biological processes. SiFT applies to a wide range of tasks, from the removal of unwanted variation in the data to revealing hidden biological structures. We demonstrate how SiFT enhances the liver circadian signal by filtering spatial zonation, recovers regenerative cell subpopulations in spatially-resolved liver data, and exposes COVID-19 disease-related cells, pathways, and dynamics by filtering healthy reference signals. SiFT performs the correction at the gene expression level, can scale to large datasets, and compares favorably to state-of-the-art methods.

Identification of hub genes and transcription factor regulatory network for heart failure using RNA-seq data and robust rank aggregation analysis.

Heart failure (HF) is the end stage of various cardiovascular diseases with a high mortality rate. Novel diagnostic and therapeutic biomarkers for HF are urgently required. Our research aims to identify HF-related hub genes and regulatory networks using bioinformatics and validation assays. Using four RNA-seq datasets in the Gene Expression Omnibus (GEO) database, we screened differentially expressed genes (DEGs) of HF using Removal of Unwanted Variation from RNA-seq data (RUVSeq) and the robust rank aggregation (RRA) method. Then, hub genes were recognized using the STRING database and Cytoscape software with cytoHubba plug-in. Furthermore, reliable hub genes were validated by the GEO microarray datasets and quantitative reverse transcription polymerase chain reaction (qRT-PCR) using heart tissues from patients with HF and non-failing donors (NFDs). In addition, R packages "clusterProfiler" and "GSVA" were utilized for enrichment analysis. Moreover, the transcription factor (TF)-DEG regulatory network was constructed by Cytoscape and verified in a microarray dataset. A total of 201 robust DEGs were identified in patients with HF and NFDs. STRING and Cytoscape analysis recognized six hub genes, among which ASPN, COL1A1, and FMOD were confirmed as reliable hub genes through microarray datasets and qRT-PCR validation. Functional analysis showed that the DEGs and hub genes were enriched in T-cell-mediated immune response and myocardial glucose metabolism, which were closely associated with myocardial fibrosis. In addition, the TF-DEG regulatory network was constructed, and 13 significant TF-DEG pairs were finally identified. Our study integrated different RNA-seq datasets using RUVSeq and the RRA method and identified ASPN, COL1A1, and FMOD as potential diagnostic biomarkers for HF. The results provide new insights into the underlying mechanisms and effective treatments of HF.

Assessing normalization methods in mass spectrometry-based proteome profiling of clinical samples

BackgroundLarge-scale proteomic studies have to deal with unwanted variability, especially when samples originate from different centers and multiple analytical batches are needed. Such variability is typically added throughout all the steps of a clinical research study, from human biological sample collection and storage, sample preparation, spectral data acquisition, to peptide and protein quantification. In order to remove such diverse and unwanted variability, normalization of the protein data is performed. There have been already several published reviews comparing normalization methods in the -omics field, but reports focusing on proteomic data generated with mass spectrometry (MS) are much fewer. Additionally, most of these reports have only dealt with small datasets. ResultsAs a case study, here we focused on the normalization of a large MS-based proteomic dataset obtained from an overweight and obese pan-European cohort, where different normalization methods were evaluated, namely: center standardize, quantile protein, quantile sample, global standardization, ComBat, median centering, mean centering, single standard and removal of unwanted variation (RUV); some of these are generic normalization methods while others have been specifically created to deal with genomic or metabolomic data. We checked how relationships between proteins and clinical variables (e.g., gender, levels of triglycerides or cholesterol) were improved after normalizing the data with the different methods. ConclusionsSome normalization methods were better adapted for this particular large-scale shotgun proteomic dataset of human plasma samples labeled with isobaric tags and analyzed with liquid chromatography-tandem MS. In particular, quantile sample normalization, RUV, mean and median centering showed very good performances, while quantile protein normalization provided worse results than those obtained with unnormalized data.

A hierarchical approach to removal of unwanted variation for large-scale metabolomics data

Liquid chromatography-mass spectrometry-based metabolomics studies are increasingly applied to large population cohorts, which run for several weeks or even years in data acquisition. This inevitably introduces unwanted intra- and inter-batch variations over time that can overshadow true biological signals and thus hinder potential biological discoveries. To date, normalisation approaches have struggled to mitigate the variability introduced by technical factors whilst preserving biological variance, especially for protracted acquisitions. Here, we propose a study design framework with an arrangement for embedding biological sample replicates to quantify variance within and between batches and a workflow that uses these replicates to remove unwanted variation in a hierarchical manner (hRUV). We use this design to produce a dataset of more than 1000 human plasma samples run over an extended period of time. We demonstrate significant improvement of hRUV over existing methods in preserving biological signals whilst removing unwanted variation for large scale metabolomics studies. Our tools not only provide a strategy for large scale data normalisation, but also provides guidance on the design strategy for large omics studies.

Exploring high-dimensional biological data with sparse contrastive principal component analysis.

Statistical analyses of high-throughput sequencing data have re-shaped the biological sciences. In spite of myriad advances, recovering interpretable biological signal from data corrupted by technical noise remains a prevalent open problem. Several classes of procedures, among them classical dimensionality reduction techniques and others incorporating subject-matter knowledge, have provided effective advances. However, no procedure currently satisfies the dual objectives of recovering stable and relevant features simultaneously. Inspired by recent proposals for making use of control data in the removal of unwanted variation, we propose a variant of principal component analysis (PCA), sparse contrastive PCA that extracts sparse, stable, interpretable and relevant biological signal. The new methodology is compared to competing dimensionality reduction approaches through a simulation study and via analyses of several publicly available protein expression, microarray gene expression and single-cell transcriptome sequencing datasets. A free and open-source software implementation of the methodology, the scPCA R package, is made available via the Bioconductor Project. Code for all analyses presented in this article is also available via GitHub. philippe_boileau@berkeley.edu. Supplementary data are available at Bioinformatics online.

ConsensusDE: an R package for assessing consensus of multiple RNA-seq algorithms with RUV correction.

Extensive evaluation of RNA-seq methods have demonstrated that no single algorithm consistently outperforms all others. Removal of unwanted variation (RUV) has also been proposed as a method for stabilizing differential expression (DE) results. Despite this, it remains a challenge to run multiple RNA-seq algorithms to identify significant differences common to multiple algorithms, whilst also integrating and assessing the impact of RUV into all algorithms. consensusDE was developed to automate the process of identifying significant DE by combining the results from multiple algorithms with minimal user input and with the option to automatically integrate RUV. consensusDE only requires a table describing the sample groups, a directory containing BAM files or preprocessed count tables and an optional transcript database for annotation. It supports merging of technical replicates, paired analyses and outputs a compendium of plots to guide the user in subsequent analyses. Herein, we assess the ability of RUV to improve DE stability when combined with multiple algorithms and between algorithms, through application to real and simulated data. We find that, although RUV increased fold change stability between algorithms, it demonstrated improved FDR in a setting of low replication for the intersect, the effect was algorithm specific and diminished with increased replication, reinforcing increased replication for recovery of true DE genes. We finish by offering some rules and considerations for the application of RUV in a consensus-based setting. consensusDE is freely available, implemented in R and available as a Bioconductor package, under the GPL-3 license, along with a comprehensive vignette describing functionality: http://bioconductor.org/packages/consensusDE/.

Batch-normalization of cerebellar and medulloblastoma gene expression datasets utilizing empirically defined negative control genes

MotivationMedulloblastoma (MB) is a brain cancer predominantly arising in children. Roughly 70% of patients are cured today, but survivors often suffer from severe sequelae. MB has been extensively studied by molecular profiling, but often in small and scattered cohorts. To improve cure rates and reduce treatment side effects, accurate integration of such data to increase analytical power will be important, if not essential.ResultsWe have integrated 23 transcription datasets, spanning 1350 MB and 291 normal brain samples. To remove batch effects, we combined the Removal of Unwanted Variation (RUV) method with a novel pipeline for determining empirical negative control genes and a panel of metrics to evaluate normalization performance. The documented approach enabled the removal of a majority of batch effects, producing a large-scale, integrative dataset of MB and cerebellar expression data. The proposed strategy will be broadly applicable for accurate integration of data and incorporation of normal reference samples for studies of various diseases. We hope that the integrated dataset will improve current research in the field of MB by allowing more large-scale gene expression analyses.Availability and implementationThe RUV-normalized expression data is available through the Gene Expression Omnibus (GEO; https://www.ncbi.nlm.nih.gov/geo/) and can be accessed via the GSE series number GSE124814.Supplementary information Supplementary data are available at Bioinformatics online.

NormalizeMets: assessing, selecting and implementing statistical methods for normalizing metabolomics data.

In metabolomics studies, unwanted variation inevitably arises from various sources. Normalization, that is the removal of unwanted variation, is an essential step in the statistical analysis of metabolomics data. However, metabolomics normalization is often considered an imprecise science due to the diverse sources of variation and the availability of a number of alternative strategies that may be implemented. We highlight the need for comparative evaluation of different normalization methods and present software strategies to help ease this task for both data-oriented and biological researchers. We present NormalizeMets-a joint graphical user interface within the familiar Microsoft Excel and freely-available R software for comparative evaluation of different normalization methods. The NormalizeMets R package along with the vignette describing the workflow can be downloaded from https://cran.r-project.org/web/packages/NormalizeMets/ . The Excel Interface and the Excel user guide are available on https://metabolomicstats.github.io/ExNormalizeMets . NormalizeMets allows for comparative evaluation of normalization methods using criteria that depend on the given dataset and the ultimate research question. Hence it guides researchers to assess, select and implement a suitable normalization method using either the familiar Microsoft Excel and/or freely-available R software. In addition, the package can be used for visualisation of metabolomics data using interactive graphical displays and to obtain end statistical results for clustering, classification, biomarker identification adjusting for confounding variables, and correlation analysis. NormalizeMets is designed for comparative evaluation of normalization methods, and can also be used to obtain end statistical results. The use of freely-available R software offers an attractive proposition for programming-oriented researchers, and the Excel interface offers a familiar alternative to most biological researchers. The package handles the data locally in the user's own computer allowing for reproducible code to be stored locally.

Removal of unwanted variation reveals novel patterns of gene expression linked to sleep homeostasis in murine cortex.

BackgroundWhy we sleep is still one of the most perplexing mysteries in biology. Strong evidence indicates that sleep is necessary for normal brain function and that sleep need is a tightly regulated process. Surprisingly, molecular mechanisms that determine sleep need are incompletely described. Moreover, very little is known about transcriptional changes that specifically accompany the accumulation and discharge of sleep need. Several studies have characterized differential gene expression changes following sleep deprivation. Much less is known, however, about changes in gene expression during the compensatory response to sleep deprivation (i.e. recovery sleep).ResultsIn this study we present a comprehensive analysis of the effects of sleep deprivation and subsequent recovery sleep on gene expression in the mouse cortex. We used a non-traditional analytical method for normalization of genome-wide gene expression data, Removal of Unwanted Variation (RUV). RUV improves detection of differential gene expression following sleep deprivation. We also show that RUV normalization is crucial to the discovery of differentially expressed genes associated with recovery sleep. Our analysis indicates that the majority of transcripts upregulated by sleep deprivation require 6 h of recovery sleep to return to baseline levels, while the majority of downregulated transcripts return to baseline levels within 1–3 h. We also find that transcripts that change rapidly during recovery (i.e. within 3 h) do so on average with a time constant that is similar to the time constant for the discharge of sleep need.ConclusionsWe demonstrate that proper data normalization is essential to identify changes in gene expression that are specifically linked to sleep deprivation and recovery sleep. Our results provide the first evidence that recovery sleep is comprised of two waves of transcriptional regulation that occur at different times and affect functionally distinct classes of genes.Electronic supplementary materialThe online version of this article (doi:10.1186/s12864-016-3065-8) contains supplementary material, which is available to authorized users.

MissMethyl: an R package for analyzing data from Illumina's HumanMethylation450 platform.

DNA methylation is one of the most commonly studied epigenetic modifications due to its role in both disease and development. The Illumina HumanMethylation450 BeadChip is a cost-effective way to profile >450 000 CpGs across the human genome, making it a popular platform for profiling DNA methylation. Here we introduce missMethyl, an R package with a suite of tools for performing normalization, removal of unwanted variation in differential methylation analysis, differential variability testing and gene set analysis for the 450K array. missMethyl is an R package available from the Bioconductor project at www.bioconductor.org. alicia.oshlack@mcri.edu.au Supplementary data are available at Bioinformatics online.

Systematic noise degrades gene co-expression signals but can be corrected

BackgroundIn the past decade, the identification of gene co-expression has become a routine part of the analysis of high-dimensional microarray data. Gene co-expression, which is mostly detected via the Pearson correlation coefficient, has played an important role in the discovery of molecular pathways and networks. Unfortunately, the presence of systematic noise in high-dimensional microarray datasets corrupts estimates of gene co-expression. Removing systematic noise from microarray data is therefore crucial. Many cleaning approaches for microarray data exist, however these methods are aimed towards improving differential expression analysis and their performances have been primarily tested for this application. To our knowledge, the performances of these approaches have never been systematically compared in the context of gene co-expression estimation.ResultsUsing simulations we demonstrate that standard cleaning procedures, such as background correction and quantile normalization, fail to adequately remove systematic noise that affects gene co-expression and at times further degrade true gene co-expression. Instead we show that a global version of removal of unwanted variation (RUV), a data-driven approach, removes systematic noise but also allows the estimation of the true underlying gene-gene correlations. We compare the performance of all noise removal methods when applied to five large published datasets on gene expression in the human brain. RUV retrieves the highest gene co-expression values for sets of genes known to interact, but also provides the greatest consistency across all five datasets. We apply the method to prioritize epileptic encephalopathy candidate genes.ConclusionsOur work raises serious concerns about the quality of many published gene co-expression analyses. RUV provides an efficient and flexible way to remove systematic noise from high-dimensional microarray datasets when the objective is gene co-expression analysis. The RUV method as applicable in the context of gene-gene correlation estimation is available as a BioconductoR-package: RUVcorr.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-015-0745-3) contains supplementary material, which is available to authorized users.