OpenAnnotate: a web server to annotate the chromatin accessibility of genomic regions

Jiang

2020

BMC Bioinformatics

Self Cite

Background In recent years, the rapid development of single-cell RNA-sequencing (scRNA-seq) techniques enables the quantitative characterization of cell types at a single-cell resolution. With the explosive growth of the number of cells profiled in individual scRNA-seq experiments, there is a demand for novel computational methods for classifying newly-generated scRNA-seq data onto annotated labels. Although several methods have recently been proposed for the cell-type classification of single-cell transcriptomic data, such limitations as inadequate accuracy, inferior robustness, and low stability greatly limit their wide applications. Results We propose a novel ensemble approach, named EnClaSC, for accurate and robust cell-type classification of single-cell transcriptomic data. Through comprehensive validation experiments, we demonstrate that EnClaSC can not only be applied to the self-projection within a specific dataset and the cell-type classification across different datasets, but also scale up well to various data dimensionality and different data sparsity. We further illustrate the ability of EnClaSC to effectively make cross-species classification, which may shed light on the studies in correlation of different species. EnClaSC is freely available at https://github.com/xy-chen16/EnClaSC. Conclusions EnClaSC enables highly accurate and robust cell-type classification of single-cell transcriptomic data via an ensemble learning method. We expect to see wide applications of our method to not only transcriptome studies, but also the classification of more general data.

Section: Discussionmentioning

confidence: 99%

EnClaSC: a novel ensemble approach for accurate and robust cell-type classification of single-cell transcriptomes

Jiang

2020

BMC Bioinformatics

Self Cite

“…The implementation of RA3 requires matched regions/features in the target single-cell data and the reference data. The web-based tool OPENANNO 29 provides a convenient way to construct the reference data: the input for OPENANNO is the peak information in single-cell data, and OPENANNO will calculate the accessibility of these peaks in 871 bulk DNase-seq samples of diverse biological context collected from ENCODE, which can be used as the reference data. Note that this approach requires the BAM files for the reference samples to calculate accessibility, an alternative approach that does not require BAM files will be discussed later.…”

Section: Ra3 Builds Effective Reference From Massive Bulk Datamentioning

confidence: 99%

“…The webserver OPENANNO 29 provides a convenient and straightforward approach to construct the reference. OPENANNO can annotate chromatin accessibility of arbitrary genomic regions by the normalized number of reads that fall into the regions using BAM files, or the normalized number of peaks that overlap with the regions using BED files.…”

Section: Openannomentioning

confidence: 99%

A reference-guided approach for epigenetic characterization of single cells

Yan

Zhang

et al. 2020

Preprint

Self Cite

1The recent advancements in single-cell technologies, including single-cell chromatin accessibility sequencing 2 (scCAS), have enabled profiling the epigenetic landscapes for thousands of individual cells. However, the character-3 istics of scCAS data, including high dimensionality, high degree of sparsity and high technical variation, make the 4 computational analysis challenging. Reference-guided approach, which utilizes the information in existing datasets, 5 may facilitate the analysis of scCAS data. We present RA3 (Reference-guided Approach for the Analysis of single-6 cell chromatin Acessibility data), which utilizes the information in massive existing bulk chromatin accessibility and 7 annotated scCAS data. RA3 simultaneously models 1) the shared biological variation among scCAS data and the 8 reference data, and 2) the unique biological variation in scCAS data that identifies distinct subpopulations. We show 9 that RA3 achieves superior performance in many scCAS datasets. We also present several approaches to construct 10 the reference data to demonstrate the wide applicability of RA3. 12Chromatin accessibility is a measure of the physical access of nuclear macromolecules to DNA and is essential for 13 understanding the regulatory mechanism 1, 2 . For rapid and sensitive probing of chromatin accessibility, assay for 14 transposase-accessible chromatin using sequencing (ATAC-seq) directly inserts sequencing adaptors into accessible 15 chromatin regions using hyperactive Tn5 transposase in vitro 3 . With the recent advancements in technology, single-cell 16 chromatin accessibility sequencing (scCAS) further enables the investigation of epigenomic landscape in individual 17 cells 4, 5 . However, the analysis of scCAS data is challenging because of its high dimensionality and high degree of 18 sparsity, as the low copy number (two of a diploid-genome) of DNA leads to only 1-10% capture rate for the hundreds 19 of thousands of possible accessible peaks 6 . The proposed approaches for the analysis of single-cell Seq) data thus present limitations due to the novelty and assay-specific challenges of extreme sparsity and tens of times 21 higher dimensions 6 . 22Several computational algorithms have been proposed to analyze scCAS data. chromVAR assesses the variation of 23 chromatin accessibility using groups of peaks that share the same functional annotations 7 . scABC calculates weights 24 of cells based on the number of distinct reads within the peak background and then uses weighted k-medoids to cluster 25 the cells 8 . cisTopic applies latent Dirichlet allocation (LDA) model to explore cis-regulatory regions and character-26 izes cell heterogeneity from the generated regions-by-topics and topics-by-cells probability matrices 9 . Cusanovich et 27 1 al. proposed a method that performs the term frequency-inverse document frequency transformation (TF-IDF) and 28 singular value decomposition (SVD) iteratively to get the final feature matrix 5, 10 . Scasat uses Jaccard distance to 29 evaluate the dissimilarity o...

Section: Enclasc Enables Cross-species Classificationmentioning

confidence: 99%

EnClaSC: A novel ensemble approach for accurate and robust cell-type classification of single-cell transcriptomes

Jiang

2019

Preprint

Self Cite

BackgroundIn recent years, the rapid development of single-cell RNA-sequencing (scRNA-seq) techniques enables the quantitative characterization of cell types at a single-cell resolution. With the explosive growth of the number of cells profiled in individual scRNA-seq experiments, there is a demand for novel computational methods for classifying newly-generated scRNA-seq data onto annotated labels. Although several methods have recently been proposed for the cell-type classification of single-cell transcriptomic data, such limitations as inadequate accuracy, inferior robustness, and low stability greatly limit their wide applications. ResultsWe propose a novel ensemble approach, named EnClaSC, for accurate and robust celltype classification of single-cell transcriptomic data. Through comprehensive validation experiments, we demonstrate that EnClaSC can not only be applied to the self-projection within a specific dataset and the cell-type classification across different datasets, but also scale up well to various data dimensionality and different data sparsity.We further illustrate the ability of EnClaSC to effectively make cross-species classification, which may shed light on the studies in correlation of different species.EnClaSC is freely available at https://github.com/xy-chen16/EnClaSC. ConclusionsEnClaSC enables highly accurate and robust cell-type classification of single-cell transcriptomic data via an ensemble learning method. We expect to see wide applications of our method to not only transcriptome studies, but also the classification of more general data.