Unsupervised clustering and epigenetic classification of single cells

Abstract

Characterizing epigenetic heterogeneity at the cellular level is a critical problem in the modern genomics era. Assays such as single cell ATAC-seq (scATAC-seq) offer an opportunity to interrogate cellular level epigenetic heterogeneity through patterns of variability in open chromatin. However, these assays exhibit technical variability that complicates clear classification and cell type identification in heterogeneous populations. We present scABC, an R package for the unsupervised clustering of single cell epigenetic data, to classify scATAC-seq data and discover regions of open chromatin specific to cell identity.Recent advances in single cell technologies such as scATAC-seq 1, 2 and scChIP-seq 3 have expanded our understanding of epigenetic heterogeneity at the single cell level. However, datasets arising from such technologies are difficult to analyze due to the inherent sparsity. In particular, consider scATAC-seq, designed to interrogate open chromatin in single cells. Open sites in a diploid genome have at most 2 chances to be captured through the assay and only a few thousand distinct reads are generated per cells, resulting in a very low chance that a particular site is captured by the assay. Consequently, it is difficult to determine whether a region is absent in an individual cell due to the lack of openness or due to the sparse nature of data. This creates a challenging task in delineating distinct subpopulations, as only a few genomic regions will have overlapping reads in a large number of cells. To avoid this issue, many studies perform FACS sorting to identify subpopulations, followed by bulk sequencing to determine genomic regions of interest and guide the single cell analysis. If the population is unknown or marker genes are unavailable, then sub-population specific analysis becomes impractical with these techniques.To combat these challenges and allow for the de novo classification of individual cells by their epigenetic signatures, we present a statistical method for the unsupervised clustering of scATAC-seq data, named scABC (single cell Accessibility Based Clustering). In contrast to previous works 2, 4 that demand predefined accessible chromatin sites, our procedure relies solely on the patterns of read counts within genomic regions to cluster cells. It requires two inputs: the individual single cell mapped read files and the full set of called peaks (which can be obtained from the union of all of the individual cells without the need for additional experiments). We apply our method to publicly available scATAC-seq data 1, 2, 4 as well as a true biological mixture to show that our approach can cluster cells with similar epigenetic patterns and identify accessible regions specific to each cluster. We further demonstrate that the cluster specific accessible regions determined by scABC have functional meaning and are capable of determining cellular identity. In particular, we show that these cluster specific accessible regions are enriched for transcription factor motifs known to be...

scBFA: modeling detection patterns to mitigate technical noise in large-scale single cell genomics data

“…We compared scBFA against PCA, Binary PCA, Scasat 34 , Destin 35 and scABC 36 , Scasat and Destin are scATAC-seq analysis tools primarily designed to identify cell types and differential accessibility analysis. Both methods treat dimensionality reduction as prior step before further clustering distinct cell types.…”

“…We performed dimensionality reduction and cell type classification experiments on several scATAC-seq datasets, analogous to our scRNA-seq analyses above. We benchmarked scBFA against PCA, Binary PCA, Scasat 34 , Destin 35 and scABC 36 . scBFA systematically outperformed all other methods in our benchmark datasets ( Fig.…”

APEC: an accesson-based method for single-cell chromatin accessibility analysis

“…To test the performance of APEC, we first obtained data from previous publications that performed scATAC-seq on a variety of cell types with known identity during hematopoietic stem cell (HSC) differentiation 20 as a gold standard. Compared to other state-of-the-art single cell chromatin accessibility analysis methods such as chromVAR 17,21 , LSI 11,12 , Cicero 18 and cisTopic 19 , this new accesson-based algorithm more precisely and clearly clustered cells into their corresponding identities according to the Adjusted Rand Index (ARI) ( Fig. 1b-c).…”

Single-cell ATAC-seq Signal Extraction and Enhancement with SCATE

“…Destin [21] applies weighted principal components and K-means clustering to the binary accessibility matrix to cluster cells. scABC [22] uses the read count matrix to cluster cells via a weighted K-medoids clustering algorithm. PRISM [23] uses the binary accessibility matrix to compute cosine distance between cells and then uses this distance to evaluate the degree of heterogeneity of a cell population.…”