EpiScanpy: integrated single-cell epigenomic analysis

Abstract

EpiScanpy is a toolkit for the analysis of single-cell epigenomic data, namely single-cell DNA methylation and single-cell ATAC-seq data. To address the modality specific challenges from epigenomics data, epiScanpy quantifies the epigenome using multiple feature space constructions and builds a nearest neighbour graph using epigenomic distance between cells. EpiScanpy makes the many existing scRNA-seq workflows from scanpy available to large-scale single-cell data from other -omics modalities, including methods for common clustering, dimension reduction, cell type identification and trajectory learning techniques, as well as an atlas integration tool for scATAC-seq datasets. The toolkit also features numerous useful downstream functions, such as differential methylation and differential openness calling, mapping epigenomic features of interest to their nearest gene, or constructing gene activity matrices using chromatin openness. We successfully benchmark epiScanpy against other scATAC-seq analysis tools and show its outperformance at discriminating cell types.

Decision level integration of unimodal and multimodal single cell data with scTriangulate

“…To determine whether scTriangulate multimodal integrated results: (1) correspond to well-described cell states, (2) have improved accuracy over alternative approaches, and (3) reveal new discrete cell populations, we first applied it to several independent human immune single-cell datasets assayed with four distinct approaches: scRNA-Seq (RNA), CITE-Seq (ADT + RNA), multiome (ATAC + RNA) and TEA-Seq (ADT + ATAC + RNA). For the analysis of snATAC-Seq, scTriangulate adopts a modified version of epiScanpy 17 to collect peak-level information for the ATAC cell clusters. For ADTs, scTriangulate uses Centered Log Ratio (CLR) normalization 1 .…”

“…We conducted QC based on both RNA and ATAC peaks. We filtered out nuclei with min_genes < 300, min_counts < 500, pct_counts_mt > 20% for RNA data, together with the additional criteria for at least 1000 peaks/nucleus in the ATAC data based on episcanpy 17 tutorial. Taken together, 10,991 nuclei were kept for further analysis.…”

ASTER: accurately estimating the number of cell types in single-cell chromatin accessibility data

“…Louvain and Leiden, which require a resolution parameter but not the number of clusters. To obtain the desired number of clusters, a binary search strategy is usually adopted ( Chen et al , 2019 , 2021 ; Danese et al , 2021 ). However, each attempt in the search process is time-consuming, especially for large data.…”

“…Although almost all the widely-used scCAS data analysis workflows, e.g. Signac ( Stuart et al , 2021 ), ArchR ( Granja et al , 2021 ) and EpiScanpy ( Danese et al , 2021 ), adopted community detection-based techniques to find the best possible grouping, the estimation of the number of cell types in scCAS data is still typically subjective and largely relied on the investigator’s desired clustering resolution and/or prior knowledge ( Supplementary Text S2 ).…”

Structure Primed Embedding on the Transcription Factor Manifold Enables Transparent Model Architectures for Gene Regulatory Network and Latent Activity Inference

“…In short, the RNA-seq data is preprocessed as detailed in Section 4.7.1, with the additional filtering of cells with > 25000 or < 1000 counts and < 20% mitocondrial counts of total. For the ATACseq data we used epiScanpy [73], filtereing peaks in < 10 cells and cells with < 5000 or > 7 • 10 4 counts, and with a variability score < 0.515. Final data contains 10411 cells, 21601 genes and 75111 peaks, Cell types are annotated using the reference PBMC dataset [58] passed to scanpy's [74] inject label transfer function, resulting in 8 annotated celltypes (Figure 5A).…”