INVESTIGATION OF CO-ACCESSIBILITY IN SINGLE CELL ATAC-SEQ AND ITS APPLICATION

One of the prevailing single cell methods is the single cell assay for transposase-accessible chromatin using sequencing (scATAC-seq). As a genome contains only a limited number of alleles (e.g. 2 alleles for diploid), the detected signal for specific loci could only range from 0, 1, or 2. This results in inherently sparse data for scATAC-seq. Recently, efforts made by researchers have largely promoted the development of analytical approaches for scATAC-seq, and these methods differ in the way the feature matrix is constructed and processed. However, in all these pipelines, the intrinsic sparsity of the data was not taken into consideration. Here, we systematically studied the pairwise correlation between signals from peaks (putative open chromatin regions) of scATAC-seq data and we asked whether merging the signals from the peaks with high correlation will reduce the sparsity of the data and improve clustering performance. We measured the co-accessibility between pairs of sites within 500 kb on the same chromosome and performed clustering using matrices constructed with peak sets of different levels of co-accessibility scores. We found that the signals from the peak sets with high co-accessibility contribute more significantly to the final clustering of cells. We merged the signals from the peak pairs with different co-accessibility score levels in a pairwise manner in our tool pcATAC (a peak coupling method for single cell ATAC-seq analysis). Here, we used the 10X PBMC scATAC-seq data as the test dataset and found that the peak coupling had little effect on clustering performance, making it a feasible method for reducing sparsity and data compression. We also tested the robustness of this method with a dataset consisting of ~2000 cells from 10 FACS-sorted cell populations in early human hematopoiesis. Quantitively the clustering performance was improved when the selection of peaks to be coupled are highly co-accessible. Overall, this study provides new insights into the analysis of scATAC-seq data.

单细胞方法中一种流行的方法是基于单细胞转座酶的染色质可及性测序 (单细胞ATAC测序)。由于基因组仅包含有限数量的等位基因（例如二倍体只有 2个等位基因），对于特定基因组位点，检测到的信号只能在 0、1 或 2 的范围内，这导致了单细胞ATAC测序数据本身的稀疏性。近来，研究者们做的许多努力在很大程度上促进了单细胞ATAC测序分析方法的发展，这些方法在构建和处理特征矩阵的方式上也大不相同。然而，在所有的这些流程中，都没有考虑数据内在的稀疏性。在这里，我们系统性地分析了单细胞ATAC测序数据中峰（推定的染色质开放区域）的区域内的信号之间的相关性，并提出是否能通过合并来自具有高相关性的峰的信号，使得我们能降低数据的稀疏性，并提高聚类效果。我们测量了同一染色体上 500kb 内的位点对之间的共可及性，并把这些位点根据共访问性得分水平进行排序，用不同得分水平区间的位点所构建的矩阵进行聚类。我们发现，来自具有高共可及性的峰的信号对细胞的最终聚类有更大的贡献。我们用我们的工具 pcATAC（一种基于耦合峰的用于单细胞ATAC数据分析的方法）以成对的方式将来自具有不同共可及性得分水平的峰的信号合并。在这里，我们使用10X PBMC单细胞ATAC测序数据作为测试数据，并发现峰的耦合对于聚类效果的影响很小，这使得这种办法有机会应用于降低稀疏性和数据压缩。我们也使用了一个由约 2000 个细胞组成、包含通过流式细胞分选得到的10种在人类造血细胞早期过程中的细胞的数据集，来测试了该方法的稳健性。在定量分析中，当选择了具有高共可及性的峰去耦合时，聚类效果有所提升。总体来说，这项研究为分析单细胞ATAC测序数据提供了新的见解。

[1] WITTKOPP P J, HAERUM B K, CLARK A G. Evolutionary changes in cis andtrans gene regulation [J]. Nature, 2004, 430(6995): 85-8.
[2] BUENROSTRO J D, GIRESI P G, ZABA L C, et al. Transposition of nativechromatin for fast and sensitive epigenomic profiling of open chromatin, DNAbindingproteins and nucleosome position [J]. Nat Methods, 2013, 10(12): 1213-8.
[3] BUENROSTRO J D, WU B, CHANG H Y, et al. ATAC-seq: A Method forAssaying Chromatin Accessibility Genome-Wide [J]. Curr Protoc Mol Biol, 2015,109: 21 29 1- 29 9.
[4] SONG L, CRAWFORD G E. DNase-seq: a high-resolution technique for mappingactive gene regulatory elements across the genome from mammalian cells [J]. ColdSpring Harb Protoc, 2010, 2010(2): pdb prot5384.
[5] THURMAN R E, RYNES E, HUMBERT R, et al. The accessible chromatinlandscape of the human genome [J]. Nature, 2012, 489(7414): 75-82.
[6] BOYLE A P, DAVIS S, SHULHA H P, et al. High-resolution mapping andcharacterization of open chromatin across the genome [J]. Cell, 2008, 132(2): 311-22.
[7] GONTARZ P, FU S H, XING X Y, et al. Comparison of differential accessibilityanalysis strategies for ATAC-seq data [J]. Sci Rep-Uk, 2020, 10(1).
[8] NEPH S, VIERSTRA J, STERGACHIS A B, et al. An expansive human regulatorylexicon encoded in transcription factor footprints [J]. Nature, 2012, 489(7414): 83-90.
[9] GUSMAO E G, ALLHOFF M, ZENKE M, et al. Analysis of computationalfootprinting methods for DNase sequencing experiments [J]. Nat Methods, 2016,13(4): 303-9.
[10] PIQUE-REGI R, DEGNER J F, PAI A A, et al. Accurate inference of transcriptionfactor binding from DNA sequence and chromatin accessibilit y data [J]. GenomeRes, 2011, 21(3): 447-55.
[11] GENOMES PROJECT C, ABECASIS G R, AUTON A, et al. An integrated map ofgenetic variation from 1,092 human genomes [J]. Nature, 2012, 491(7422): 56-65.
[12] ROADMAP EPIGENOMICS C, KUNDAJE A, MEULEMAN W, et al . Integrativeanalysis of 111 reference human epigenomes [J]. Nature, 2015, 518(7539): 317-30.
[13] ACKERMANN A M, WANG Z, SCHUG J, et al. Integration of ATAC-seq andRNA-seq identifies human alpha cell and beta cell signature genes [J]. Mol Metab,2016, 5(3): 233-44.􀀵􀀨􀀩􀀨􀀵􀀨􀀱􀀦􀀨􀀶􀀚􀀕
[14] YAN F, POWELL D R, CURTIS D J, et al. From reads to insight: a hitchhiker'sguide to ATAC-seq data analysis [J]. Genome Biol, 2020, 21(1): 22.
[15] FastQC [M]. Andrews S. Babraham Bioinformatics. 2015.
[16] MARTIN M. Cutadapt removes adapter sequences from high-throughput sequencingreads [J]. EMBnet J, 2011, 17(1): 10-12.
[17] BOLGER A M, LOHSE M, USADEL B. Trimmomatic: a flexible trimmer forIllumina sequence data [J]. Bioinformatics, 2014, 30(15): 2114-20.
[18] LI H, DURBIN R. Fast and accurate short read alignment with Burrows-Wheelertransform [J]. Bioinformatics, 2009, 25(14): 1754-60.
[19] LANGMEAD B, SALZBERG S L. Fast gapped-read alignment with Bowtie 2 [J].Nat Methods, 2012, 9(4): 357-9.
[20] ZHANG Y, LIU T, MEYER C A, et al. Model-based analysis of ChIP-Seq (MACS)[J]. Genome Biol, 2008, 9(9): R137.
[21] KIM T H, REN B. Genome-wide analysis of protein-DNA interactions [J]. AnnuRev Genom Hum G, 2006, 7: 81-102.
[22] HEINZ S, BENNER C, SPANN N, et al. Simple combinations of lineagedeterminingtranscription factors prime cis-regulatory elements required formacrophage and B cell identities [J]. Mol Cell, 2010, 38(4): 576-89.
[23] TARBELL E D, LIU T. HMMRATAC: a Hidden Markov ModeleR for ATAC-seq[J]. Nucleic Acids Res, 2019, 47(16): e91.
[24] YU G, WANG L G, HE Q Y. ChIPseeker: an R/Bioconductor package for ChIPpeak annotation, comparison and visualization [J]. Bioinformatics, 2015, 31(14):2382-3.
[25] ZHU L J, GAZIN C, LAWSON N D, et al. ChIPpeakAnno: a Bioconductor packageto annotate ChIP-seq and ChIP-chip data [J]. Bmc Bioinformatics, 2010, 11.
[26] GRANT C E, BAILEY T L, NOBLE W S. FIMO: scanning for occurrences of agiven motif [J]. Bioinformatics, 2011, 27(7): 1017-8.
[27] KHAN A, FORNES O, STIGLIANI A, et al. JASPAR 2018: update of the openaccessdatabase of transcription factor binding profiles and its web framework [J].Nucleic Acids Res, 2018, 46(D1): D1284.
[28] MATYS V, KEL-MARGOULIS O V, FRICKE E, et al. TRANSFAC (R) and itsmodule TRANSCompel (R): transcriptional gene regulation in eukaryotes [J].Nucleic Acids Research, 2006, 34(D108-D10.
[29] KULAKOVSKIY I V, VORONTSOV I E, YEVSHIN I S, et al. HOCOMOCO:towards a complete collection of transcription factor binding models for human andmouse via large-scale ChIP-Seq analysis [J]. Nucleic Acids Res, 2018, 46(D1):D252-D9.􀀵􀀨􀀩􀀨􀀵􀀨􀀱􀀦􀀨􀀶􀀚􀀖
[30] ROBINSON M D, MCCARTHY D J, SMYTH G K. edgeR: a Bioconductor packagefor differential expression analysis of digital gene expression data [J].Bioinformatics, 2010, 26(1): 139-40.
[31] LOVE M I, HUBER W, ANDERS S. Moderated estimation of fold change anddispersion for RNA-seq data with DESeq2 [J]. Genome Biol, 2014, 15(12): 550.
[32] ANDERSSON R, SANDELIN A. Determinants of enhancer and promoter activitiesof regulatory elements [J]. Nat Rev Genet, 2020, 21(2): 71-87.
[33] GAFFNEY D J. Mapping and predicting gene-enhancer interactions [J]. Nat Genet,2019, 51(12): 1662-3.
[34] SCHOENFELDER S, FRASER P. Long-range enhancer-promoter contacts in geneexpression control [J]. Nature Reviews Genetics, 2019, 20(8): 437-55.
[35] GATE R E, CHENG C S, AIDEN A P, et al. Genetic determinants of co-accessiblechromatin regions in activated T cells across humans [J]. Nat Genet, 2018, 50(8):1140-50.
[36] KUMASAKA N, KNIGHTS A J, GAFFNEY D J. High-resolution genetic mappingof putative causal interactions between regions of open chromatin [J]. NatureGenetics, 2019, 51(1): 128-+.
[37] 􀀼􀀲􀀸􀀱􀀪􀀃􀀪􀀵􀀨􀀨􀀱􀀺􀀤􀀯􀀧􀀃􀀺􀀃􀀺􀀏􀀃􀀧􀂶􀀤􀀱􀀷􀀲􀀱􀀬􀀲-CHRONOWSKA A, BENAGLIO P, etal. Chromatin co-accessibility is highly structured, spans entire chromosomes, andmediates long range regulatory genetic effects [J]. bioRxiv, 2019, 604371.
[38] PLINER H A, PACKER J S, MCFALINE-FIGUEROA J L, et al. Cicero Predictscis-Regulatory DNA Interactions from Single-Cell Chromatin Accessibility Data [J].Mol Cell, 2018, 71(5): 858-71 e8.
[39] DONG K N, ZHANG S H. Joint reconstruction of cis-regulatory interactionnetworks across multiple tissues using single-cell chromatin accessibility data [J].Brief Bioinform, 2021, 22(3):
[40] WANG Z, GERSTEIN M, SNYDER M. RNA-Seq: a revolutionary tool fortranscriptomics [J]. Nat Rev Genet, 2009, 10(1): 57-63.
[41] STUART T, SATIJA R. Integrative single-cell analysis [J]. Nature ReviewsGenetics, 2019, 20(5): 257-72.
[42] TRAPNELL C. Defining cell types and states with single-cell genomics [J].Genome Res, 2015, 25(10): 1491-8.
[43] WAGNER D E, KLEIN A M. Lineage tracing meets single-cell omics: opportunitiesand challenges [J]. Nat Rev Genet, 2020, 21(7): 410-27.
[44] QIU X J, MAO Q, TANG Y, et al. Reversed graph embedding resolves complexsingle-cell trajectories [J]. Nature Methods, 2017, 14(10): 979-+.
[45] REGEV A, TEICHMANN S A, LANDER E S, et al. The Human Cell Atlas [J].Elife, 2017, 6.􀀵􀀨􀀩􀀨􀀵􀀨􀀱􀀦􀀨􀀶􀀚􀀗
[46] TANG F C, BARBACIORU C, WANG Y Z, et al. mRNA-Seq whole-transcriptomeanalysis of a single cell [J]. Nature Methods, 2009, 6(5): 377-U86.
[47] KLEIN A M, MAZUTIS L, AKARTUNA I, et al. Droplet Barcoding for Single -CellTranscriptomics Applied to Embryonic Stem Cells [J]. Cell, 2015, 161(5): 1187-201.
[48] CUSANOVICH D A, DAZA R, ADEY A, et al. Multiplex single cell profiling ofchromatin accessibility by combinatorial cellular indexing [J]. Science, 2015,348(6237): 910-4.
[49] CHEN X, MIRAGAIA R J, NATARAJAN K N, et al. A rapid and robust method forsingle cell chromatin accessibility profiling [J]. Nat Commun, 2018, 9(1): 5345.
[50] BUENROSTRO J D, WU B, LITZENBURGER U M, et al. Single-cell chromatinaccessibility reveals principles of regulatory variation [J]. Nature, 2015, 523(7561):486-90.
[51] MA S, ZHANG B, LAFAVE L M, et al. Chromatin Potential Identified by SharedSingle-Cell Profiling of RNA and Chromatin [J]. Cell, 2020, 183(4): 1103-16 e20.
[52] STUART T, BUTLER A, HOFFMAN P, et al. Comprehensive Integration of Single -Cell Data [J]. Cell, 2019, 177(7): 1888-902 e21.
[53] CHEN H, LAREAU C, ANDREANI T, et al. Assessment of computational methodsfor the analysis of single-cell ATAC-seq data [J]. Genome Biol, 2019, 20(1): 241.
[54] MEREU E, LAFZI A, MOUTINHO C, et al. Benchmarking single-cell RNAsequencingprotocols for cell atlas projects [J]. Nat Biotechnol, 2020, 38(6): 747-+.
[55] CUSANOVICH D A, HILL A J, AGHAMIRZAIE D, et al. A Single-Cell Atlas of InVivo Mammalian Chromatin Accessibility [J]. Cell, 2018, 174(5): 1309-24 e18.
[56] CUSANOVICH D A, REDDINGTON J P, GARFIELD D A, et al. The cisregulatorydynamics of embryonic development at single-cell resolution [J]. Nature,2018, 555(7697): 538-+.
[57] ZHAO C C, HU S E, HUO X, et al. Dr. seq2: A quality control and analysispipeline for parallel single cell transcriptome and epigenome data [J]. Plos One,2017, 12(7):
[58] JI Z, ZHOU W, JI H. Single-cell regulome data analysis by SCRAT [J].Bioinformatics, 2017, 33(18): 2930-2.
[59] SCHEP A N, WU B, BUENROSTRO J D, et al. chromVAR: inferring transcription -factor-associated accessibility from single-cell epigenomic data [J]. Nat Methods,2017, 14(10): 975-8.
[60] DE BOER C G, REGEV A. BROCKMAN: deciphering variance in epigenomicregulators by k-mer factorization [J]. BMC Bioinformatics, 2018, 19(1): 253.
[61] ZAMANIGHOMI M, LIN Z, DALEY T, et al. Unsupervised clustering andepigenetic classification of single cells [J]. Nat Commun, 2018, 9(1): 2410.􀀵􀀨􀀩􀀨􀀵􀀨􀀱􀀦􀀨􀀶􀀚􀀘
[62] BRAVO GONZALEZ-BLAS C, MINNOYE L, PAPASOKRATI D, et al. cisTopic:cis-regulatory topic modeling on single-cell ATAC-seq data [J]. Nat Methods, 2019,16(5): 397-400.
[63] LAREAU C A, DUARTE F M, CHEW J G, et al. Droplet-based combinatorialindexing for massive-scale single-cell chromatin accessibility [J]. Nat Biotechnol,2019, 37(8): 916-24.
[64] BAKER S M, ROGERSON C, HAYES A, et al. Classifying cells with Scasat, asingle-cell ATAC-seq analysis tool [J]. Nucleic Acids Res, 2019, 47(2): e10.
[65] URRUTIA E, CHEN L, ZHOU H, et al. Destin: toolkit for single-cell analysis ofchromatin accessibility [J]. Bioinformatics, 2019, 35(19): 3818-20.
[66] CAI S, GEORGAKILAS G K, JOHNSON J L, et al. A Cosine Similarity-BasedMethod to Infer Variability of Chromatin Accessibility at the Single -Cell Level [J].Front Genet, 2018, 9.
[67] YU W, UZUN Y, ZHU Q, et al. scATAC-pro: a comprehensive workbench forsingle-cell chromatin accessibility sequencing data [J]. Genome Biol, 2020, 21(1):94.
[68] TRAPNELL C, CACCHIARELLI D, GRIMSBY J, et al. The dynamics andregulators of cell fate decisions are revealed by pseudotemporal ordering of singlecells [J]. Nat Biotechnol, 2014, 32(4): 381-6.
[69] STUART T, SRIVASTAVA A, LAREAU C, et al. Multimodal single -cell chromatinanalysis with Signac [J]. bioRxiv, 2020, 2020.11.09.373613.
[70] FANG R, PREISSL S, LI Y, et al. Comprehensive analysis of single cell ATAC-seqdata with SnapATAC [J]. Nat Commun, 2021, 12(1): 1337.
[71] GRANJA J M, CORCES M R, PIERCE S E, et al. ArchR is a scalable softwarepackage for integrative single-cell chromatin accessibility analysis [J]. NatureGenetics, 2021,
[72] JI Z, ZHOU W, HOU W, et al. Single-cell ATAC-seq signal extraction andenhancement with SCATE [J]. Genome Biol, 2020, 21(1): 161.
[73] Picard toolkit [M]. Broad Institute, GitHub repository; Broad Institute. 2019.
[74] LI H, HANDSAKER B, WYSOKER A, et al. The Sequence Alignment/Map formatand SAMtools [J]. Bioinformatics, 2009, 25(16): 2078-9.
[75] AMEMIYA H M, KUNDAJE A, BOYLE A P. The ENCODE Blacklist:Identification of Problematic Regions of the Genome [J]. Sci Rep, 2019, 9(1): 9354.
[76] CONSORTIUM E P. An integrated encyclopedia of DNA elements in the humangenome [J]. Nature, 2012, 489(7414): 57-74.
[77] BERNSTEIN B E, STAMATOYANNOPOULOS J A, COSTELLO J F, et al. TheNIH Roadmap Epigenomics Mapping Consortium [J]. Nat Biotechnol, 2010, 28(10):1045-8.􀀵􀀨􀀩􀀨􀀵􀀨􀀱􀀦􀀨􀀶􀀚􀀙
[78] HAO Y, HAO S, ANDERSEN-NISSEN E, et al. Integrated analysis of multimodalsingle-cell data [J]. Cell, 2021, 184(13): 3573-87 e29.
[79] KURIMO M. Indexing audio documents by using latent semantic analysis and SOM[J]. Kohonen Maps, 1999, 363-74.
[80] GRANJA J M, CORCES M R, PIERCE S E, et al. ArchR is a scalable softwarepackage for integrative single-cell chromatin accessibility analysis (vol 53, pg 403,2021) [J]. Nature Genetics, 2021, 53(6): 935.
[81] KORSUNSKY I, MILLARD N, FAN J, et al. Fast, sensitive and accurate integrationof single-cell data with Harmony [J]. Nat Methods, 2019, 16(12): 1289-96.
[82] SATPATHY A T, GRANJA J M, YOST K E, et al. Massively parallel single-cellchromatin landscapes of human immune cell development and intratumoral T cellexhaustion [J]. Nat Biotechnol, 2019, 37(8): 925-36.
[83] GRANJA J M, KLEMM S, MCGINNIS L M, et al. Single-cell multiomic analysisidentifies regulatory programs in mixed-phenotype acute leukemia [J]. NatBiotechnol, 2019, 37(12): 1458-65.
[84] WALTMAN L, VAN ECK N J. A smart local moving algorithm for large -scalemodularity-based community detection [J]. Eur Phys J B, 2013, 86(11).
[85] MCINNES L, HEALY J, MELVILLE J. Umap: Uniform manifold approximationand projection for dimension reduction [J]. arXiv preprint arXiv:180203426, 2018.
[86] VAN DER MAATEN L, HINTON G. Visualizing Data using t -SNE [J]. J MachLearn Res, 2008, 9: 2579-605.
[87] JIANG L, CHEN H D, PINELLO L, et al. GiniClust: detecting rare cell types fromsingle-cell gene expression data with Gini index [J]. Genome Biology, 2016, 17.
[88] TSOUCAS D, YUAN G C. GiniClust2: a cluster-aware, weighted ensembleclustering method for cell-type detection [J]. Genome Biology, 2018, 19.
[89] PLINER H A, SHENDURE J, TRAPNELL C. Supervised classification enablesrapid annotation of cell atlases [J]. Nat Methods, 2019, 16(10): 983-6.
[90] GRIFFITHS J A, RICHARD A C, BACH K, et al. Detection and removal of barcodeswapping in single-cell RNA-seq data [J]. Nat Commun, 2018, 9(1): 2667.
[91] LUN A T L, RIESENFELD S, ANDREWS T, et al. EmptyDrops: distinguishingcells from empty droplets in droplet-based single-cell RNA sequencing data [J].Genome Biol, 2019, 20(1): 63.
[92] FRIEDMAN J, HASTIE T, TIBSHIRANI R. Sparse inverse covariance estimationwith the graphical lasso [J]. Biostatistics, 2008, 9(3): 432-41.
[93] JAVIERRE B M, BURREN O S, WILDER S P, et al. Lineage-Specific GenomeArchitecture Links Enhancers and Non-coding Disease Variants to Target GenePromoters [J]. Cell, 2016, 167(5): 1369-+.􀀵􀀨􀀩􀀨􀀵􀀨􀀱􀀦􀀨􀀶􀀚􀀚
[94] BUENROSTRO J D, CORCES M R, LAREAU C A, et al. Integrated Single -CellAnalysis Maps the Continuous Regulatory Landscape of Human HematopoieticDifferentiation [J]. Cell, 2018, 173(6): 1535-48 e16.
[95] WANG X, SHEN X, CHEN S, et al. Reinvestigation of Classic T Cell Subsets andIdentification of Novel Cell Subpopulations by Single-Cell RNA Sequencing [J]. JImmunol, 2022, 208(2): 396-406.
[96] TANG F, BARBACIORU C, WANG Y, et al. mRNA-Seq whole-transcriptomeanalysis of a single cell [J]. Nat Methods, 2009, 6(5): 377-82.
[97] WANG C F, SUN D Q, HUANG X, et al. Integrative analyses of single -celltranscriptome and regulome using MAESTRO [J]. Genome Biology, 2020, 21(1).
[98] STOECKIUS M, HAFEMEISTER C, STEPHENSON W, et al. Simultaneous epitopeand transcriptome measurement in single cells [J]. Nat Methods, 2017, 14(9): 865-8.
[99] JUMPER J, EVANS R, PRITZEL A, et al. Highly accurate protein structureprediction with AlphaFold [J]. Nature, 2021, 596(7873): 583-9.
[100] LI Z, LI Y, ZHANG B, et al. DeeReCT-APA: Prediction of AlternativePolyadenylation Site Usage Through Deep Learning [J]. Genomics ProteomicsBioinformatics, 2021.
[101] ZENG T, LI Y I. Predicting RNA splicing from DNA sequence using Pangolin [J].bioRxiv, 2021, 2021.07.06.451243.
[102] JAGANATHAN K, KYRIAZOPOULOU PANAGIOTOPOULOU S, MCRAE J F, etal. Predicting Splicing from Primary Sequence with Deep Learning [J]. Cell, 2019,176(3): 535-48 e24.