High-throughput single-cell joint analysis of histone modications and gene expression by Paired-Tag

We describe here Paired-Tag, a high-throughput multi-omics method for joint proling of histone modications and gene expressions in single cells. The assay is based on a combinatorial barcoding indexing strategy that does not require special instruments. It can be performed with nuclei extracted from cultured cells or frozen tissues, in standard molecular biology laboratories.


Introduction
Covalent modi cations to histone proteins play critical roles in regulating cell-type-speci c gene expression programs in multi-cellular organisms 1,2 . Sigle-cell methods for pro ling histone modi cations [3][4][5][6][7][8][9] were recently reported. However, integrative analysis of different histone modi cations in heterogeneous tissues is very challenging because of the di culty in aligning cell clusters de ned based on different histone marks, which have different relationships to gene expression and cell states. Paired-Tag addresses this challenge by enabling simultaneous pro ling of gene expression and histone modi cations from the same individual cells. Cells of the same types in a heterogenous population could be grouped together based on their similar transcription pro les in the RNA component of the Paired-Tag pro les, and the different histone modi cation pro les could be directly obtained and integrated from the DNA component of the Paired-Tag datasets.
Paired-Tag builds upon a technique that we previously developed, Paired-seq 10 , which utilizes a combinatorial barcode indexing strategy 11 for joint analysis of gene expression and chromatin accessibilities in single cells. Paired-Tag combines Paired-seq with the CUT&Tag strategy 7 . Like Pairedseq, Paired-Tag does not require speci c instruments and can be performed in standard molecular biology laboratories. It is highly-scalable, capable of processing up to 500,000 nuclei in a single experiment. It generates joint histone modi cations and gene expression pro les as effective as standalone single cell scRNA-seq and scChIP-seq assays.  5. Use Agilent Fragment Analyzer to analysis the fragment distribution of libraries (Fig. 1). If there is a clear band at ~245 bp (Fig. 1c), repeat the size selection with 0.75X SPRI beads or other size selection method (see Troubleshooting for details). D. Initial data processing and QC 1. All scripts for initial Paired-Tag data processing are available from GitHub (https://github.com/cxzhu/Paired-Tag). Download and perform the data pre-processing steps according to instructions on the GitHub repository.
2. Initial Paired-Tag data processing includes (a) extract barcode sequences from Read2, (b) assign barcodes combinations to cellular barcodes references (assign barcode sequences to ID of 12 sample tubes and 2 rounds of 96 wells), (c) mapping the assigned reads to reference genome and (d) generating cell-to-features matrices for downstream analyses.
3. The following metrics during initial Paired-Tag data processing can be used for quality control. For step 2(a), typically >85% and >75% of DNA and RNA reads will have full ligated barcodes. For step 2(b), >85% of both DNA and RNA reads can uniquely assigned to one cellular barcode with no more than 1 mismatch. For step 2(c), typically >85% of assigned reads can be mapped to the reference genome; depending on which histone mark targeted, from 60% to >95% of assigned DNA reads can be mapped to the reference genome.
4. Cell-to-features matrices generated from initial data processing steps is a standard sparse matrix format which can be used for down-stream analyses with commonly used single-cell analysis software.
Troubleshooting · After pre-ampli cation, typically ~30 -1,000 ng DNA will be recovered (around 1 ng/µL to 30 ng/µL measured by Qubit, when pre-ampli ed products were eluted in 36 µL). If the concentration is less than 0.2 ng/µL: (1) considering check the reagents used for pre-ampli cation steps (C1); (2) pre-ampli cation may be inhibited by residual SDS from B-5 step4, considering repeat this step (B-5 step4); (3) the quality of extracted nuclei may not be good enough for Paired-Tag experiment, considering perform a bulk ATACseq from extracted nuclei for QC. If the concentration is higher than 50 ng/µL, the ampli ed products are likely to be over-ampli ed barcode-adaptor dimers, please see the next paragraph for details. · If both bands of ~245 bp and 300-1000 bp exist for DNA or RNA libraries (Fig. 1c), indicating the presence of both libraries and barcode-adaptor dimers. (1) If a clear ~245 bp band exists, consider perform an additional 0.75X SPRI beads size selection (C-3 step4) (or other size selection method) to remove the primer-barcode dimers (~245 bp). (2) If there is no ~245 bp band but with signi cant fraction of fragments with length < 300 bp (e.g., Fig. 1c), consider reduce the amount of N5-Tn5 used in C-2 step8.