18 September 2020 in PLOS Computational Biology
Single-cell Hi-C (scHi-C) interrogates genome-wide chromatin interaction in individual cells, allowing us to gain insights into 3D genome organization. However, the extremely sparse nature of scHi-C data poses a significant barrier to analysis, limiting our ability to tease out hidden biological information. In this work, we approach this problem by applying topic modeling to scHi-C data. Topic modeling is well-suited for discovering latent topics in a collection of discrete data. For our analysis, we generate nine different single-cell combinatorial indexed Hi-C (sci-Hi-C) libraries from five human cell lines (GM12878, H1Esc, HFF, IMR90, and HAP1), consisting over 19,000 cells. We demonstrate that topic modeling is able to successfully capture cell type differences from sci-Hi-C data in the form of “chromatin topics.” We further show enrichment of particular compartment structures associated with locus pairs in these topics. The genomes of higher organisms are intricately folded and organized in a dynamic manner that has strong implications for many biological processes. Each chromosome undergoes dramatic changes to their three dimensional conformation during the cell cycle, whereas the positioning of chromosomes within the nucleus plays an important role in controlling the activation of specific genes. Recently, it has become possible to investigate the 3D conformations of the genomes of individual cells using a high throughput sequencing assay called single cell Hi-C (scHi-C). However, data from these assays are sparse and noisy, making analysis and interpretation of scHi-C data challenging. In this work, we generated a scHi-C dataset of over 19,000 cells from five human cell lines and applied a natural language processing method called topic modeling to discover cell type-specific “chromatin” topics. We show that these topics can be used to distinguish between cells at different stages of the cell cycle and cells from different tissues based on the 3D conformation of their genomes, despite the sparsity of the data. We further show that the 3D conformations of single cells are linked to the expression of cell type-specific genes and to cell cycle-associated conformational patterns.
6 October 2023 in bioRxiv
Copy number variation (CNV) is a widely studied type of structural variation seen in the genomes of cancerous and other dysfunctional cells. CNVs can have direct and indirect effects on gene dosage, and are thought to drive cancer progression and other disorders. Advancements in single-cell assays such as sc-ATAC-seq and sc-RNA-seq, along with their ubiquitous use, allows for the identification of CNVs at single cell resolution. While there are a variety of available tools for CNV detection in sc-RNA-seq, development of sc-ATAC-seq based accurate and reliable CNV callers is in the early stages, with only two available algorithms so far. We present RIDDLER, a single-cell ATAC-seq CNV detection algorithm based on outlier aware generalized linear modeling. By utilizing tools from robust statistics, we developed an extensible model that is able to identify single-cell CNVs from sc-ATAC-seq data in an unsupervised fashion, while providing probabilistic justification for results. Our statistical approach also allows us to estimate when loss of signal is likely caused by drop-out or a true genome deletion event, as well as predict reliable CNVs without the need for normative reference cells. We demonstrate the effectiveness of our algorithm on cancer cell line models where it achieves better agreement with bulk WGS derived CNVs than competing methods. We also compare our approach on 10x multimone data, where it shows better agreement and integration with RNA derived CNV estimates.
12 January 2021 in bioRxiv
Single-cell genomics assays have emerged as a dominant platform for interrogating complex biological systems. Methods to capture various properties at the single-cell level typically suffer a tradeoff between cell count and information content, which is defined by the number of unique and usable reads acquired per cell. We and others have described workflows that utilize single-cell combinatorial indexing (sci) 1 , leveraging transposase-based library construction 2 to assess a variety of genomic properties in high throughput; however, these techniques often produce sparse coverage for the property of interest. Here, we describe a novel adaptor-switching strategy, ‘s3’, capable of producing one-to-two order-of-magnitude improvements in usable reads obtained per cell for chromatin accessibility (s3-ATAC), whole genome sequencing (s3-WGS), and whole genome plus chromatin conformation (s3-GCC), while retaining the same high-throughput capabilities of predecessor ‘sci’ technologies. We apply s3 to produce high-coverage single-cell ATAC-seq profiles of mouse brain and human cortex tissue; and whole genome and chromatin contact maps for two low-passage patient-derived cell lines from a primary pancreatic tumor.
12 January 2022 in bioRxiv
Genome-wide mapping of the histone modification landscape is critical to understanding tran-scriptional regulation. Cleavage Under Targets and Tagmentation (CUT&Tag) is a new method for profiling the localization of covalent histone modifications, offering improved sensitivity and decreased cost compared with Chromatin Immunoprecipitation Sequencing (ChIP-seq). Here, we present GoPeaks, a peak calling method specifically designed for histone modification CUT&Tag data. GoPeaks implements a Binomial distribution and stringent read count cut-off to nominate candidate genomic regions. We compared the performance of GoPeaks against commonly used peak calling algorithms to detect H3K4me3, H3K4me1, and H3K27Ac peaks from CUT&Tag data. These histone modifications display a range of peak profiles and are frequently used in epigenetic studies. We found GoPeaks robustly detects genome-wide histone modifications and, notably, identifies H3K27Ac with improved sensitivity compared to other standard peak calling algorithms.
17 July 2026 in Nucleic Acids Research
Copy number variations (CNVs) are genomic structural variants that are strongly linked to cancer progression and genetic disorders. CNVs can be highly heterogeneous at population and tissue scale; thus, single-cell resolution detection holds great promise for studying clonal evolution and CNV-driven changes. Despite advanced sc-RNA-seq CNV detection methods, accurate methods for epigenomic single-cell modalities lag behind. We developed RIDDLER; a robust, unsupervised method that uses outlier-aware statistical modeling to detect CNVs across multiple single-cell modalities and assays. RIDDLER utilizes a robust regression framework to model the expected distribution of reads genome-wide by accounting for assay-specific biases, identifying CNVs as outliers from that distribution. This versatile framing allows deployment of RIDDLER in multiple modalities with appropriate bias features. We demonstrate the accuracy of RIDDLER in calling single-cell CNVs and dissecting clonal heterogeneity in sc-ATAC-seq and sc-methylation. RIDDLER is more accurate and more robust to data sparsity than competing methods. We illustrate useful applications of RIDDLER for dissection of clonal structure, identification of subclonal accessibility peaks, and multimodal integration from CNV structure. RIDDLER stands out as a scalable, generalizable multi-modal method for accurate CNV detection, empowering studies aiming to link CNV dynamics to epigenetic alterations within the same cell.