Optimization of the Omni-ATAC protocol to chromatin accessibility profiling in snap-frozen rat adipose and muscle tissues

ATAC-seq is a fast and sensitive method for the epigenomic profiling of open chromatin and for mapping of transcription factor binding sites [1]. Despite the development of the Omni-ATAC protocol for the profiling of chromatin accessibility in frozen tissues [2], studies in adipose tissue have been restricted due to technical challenges including the high lipid content of adipocytes and reproducibility issues between replicates. Here, we provide a modified Omni-ATAC protocol that achieves high data reproducibility in various tissue types from rat, including adipose and muscle tissues [3].• This protocol describes a methodology that enables chromatin accessibility profiling from snap-frozen rat adipose and muscle tissues.• The technique comprises an optimized bead-based tissue homogenization process that substitutes to Dounce homogenization, reduces variability in the experimental procedure, and is adaptable to various tissue types.• In comparison with the Omni-ATAC protocol, the method described here results in improved ATAC-seq data quality that complies with ENCODE quality standards.


Method Introduction
The Assay for Transposase-Accessible Chromatin using sequencing (ATAC-seq), is a low-input, relatively fast and easy-to-perform method, which consequently has been widely used for mapping chromatin accessibility genome-wide and for investigating TF binding [ 1 , 4 ]. It uses a genetically engineered hyperactive Tn5 transposase that simultaneously cleaves DNA in open chromatin regions and ligates adapters for high-throughput sequencing at these regions [ 5 , 6 ]. While the Omni-ATAC protocol was developed to profile chromatin accessibility in frozen tissues [2] , it has shown a few limitations. Notably, ATAC-seq analyses in adipose tissue or in tissue-derived adipocytes have been scarce, most likely due to the technical challenge posed by the high lipid content of adipocytes [7][8][9][10][11] . Moreover, low reproducibility between replicates has been observed, with variability in Irreproducible Discovery Rate (IDR) peaks and in transcription start site (TSS) enrichment scores across tissues [9] . In the Omni-ATAC protocol, tissues are homogenized using a Dounce homogenizer, involving a defined number of strokes with the pestles (e.g. 10 with the loose pestle, followed by 20 with the tight pestle), which may differ between tissue types (e.g. soft vs. hard tissues) [12] . Subsequently, the number of isolated nuclei can vary between tissue types, adding some sample-to-sample variability. To optimize the ATAC-seq method to frozen adipose tissue and gain high data reproducibility in various tissues, we modified the Omni-ATAC protocol to include an effective bead-based tissue homogenization process that is applicable to adipose tissue and skeletal muscle. Fig. 1 illustrates the differences between the old and the new protocol. The resulting Ruptor-ATAC protocol is suitable for high throughput studies, complies with the ENCODE quality standards for ATAC-seq data ( https://www.encodeproject. org/atac-seq/#standards ; [13] ), and demonstrates improved ATAC-seq data quality based on metrics such as Usable Reads, IDR peaks, FRiP and TSS enrichment scores.

Ruptor-ATAC protocol
Note : All steps should be performed on ice or at 4 °C. Pre-chill a centrifuge to 4 °C. Frozen tissue fragments from rat ( ∼10 mg for muscle and ∼30 mg for adipose tissue) are used in this protocol.
Required reagents and equipment:  Step 1: Tissue homogenization Note : Omni BR Cryo cooling unit is kept at 10 °C ahead of time and throughout the tissue homogenization process using liquid nitrogen.

Stock solution and buffer preparation
Prepare 4x Homogenization Stock Buffer in advance and store at 4 0 C. Sterile filtration is recommended. 4x

ATAC-seq library quality control
1. Determine the quantity of purified libraries using the Qubit dsDNA high sensitivity assay kit, which provides the DNA concentration measured on the fluorometer. 2. Assess the quality of purified libraries using an Agilent Bioanalyzer High-Sensitivity DNA kit, which gives information about the size distribution.
The nucleosome is comprised of a histone octamer that is complexed with approximately 147 bp of DNA [14] . Adapters are an additional 142 bp. A typical fragment size distribution plot shows an enrichment in nucleosome-free fragments ( ∼200 bp) and mono-nucleosome-bound fragments ( ∼300 bp), followed by di-nucleosome-associated regions ( ∼500 bp). Both adipose and muscle ATAC-seq libraries show the expected fragment size and nucleosome phasing ( Fig. 2 ).

ATAC-seq data analysis and quality control
Required software: • ENCODE ATAC-seq pipeline v1.7.0 ( https://www.encodeproject.org/atac-seq/ ) • FastQC [15] Table 1 Comparison of the ATAC-seq data quality metrics obtained using Ruptor-ATAC (see in [3] ) vs. those previously collected in mouse adipose and skeletal muscle tissue using Omni-ATAC [9] . Mapped reads, total number of reads minus number of unaligned reads. chrM reads, mitochondrial chromosome reads. Usable reads, number of mapped reads minus number of low mapping quality, duplicate, and mitochondrial reads. % Usable reads, Usable reads to Total reads ratio. TSS enrichment, TSS enrichment score. IDR Peaks, Irreproducible Discovery Rate on more than two replicates. FRiP, Fraction of reads in peaks. N.a., not available. 1. Process the ATAC-seq data (trim, align, filter, and quality control the data) using the ENCODE ATAC-seq pipeline v1.7.0. Table 1 presents the ATAC-seq metadata and mapping statistics obtained using the Ruptor-ATAC protocol, along with the ATAC-seq metrics previously acquired by Liu et al. using the Omni-ATAC protocol [9] . Reproducible peaks range from 44,673 to 54,458 in adipose, and from 84,195 to 112,272 in muscle samples (IDR peaks). TSS enrichment values are generally well above the cutoff ( > 9; https://www.encodeproject.org/atac-seq/#standards ), and the fraction of reads in called peak regions (FRiP) scores are mostly greater than 0.2, denoting the high quality of the ATAC-seq data. 2. For all samples, read quality is assessed using FastQC. 3. Trimmomatic is used to remove adapters and low-quality base pairs and reads are identified by FastQC. 4. Reads for each sample are aligned to the rat genome (rn6.0) using Bowtie2. 5. After alignment, reads mapping to the mitochondrial genome are removed. 6. As a general measure of sensitivity to Tn5 transposase fragmentation, ATAC-seq signal is defined as number of transposase cuts mapping to each bp (or a total count of cuts mapping to a selected genomic interval). The cuts are defined as 5' ends of ATAC-seq reads, with additional shifting by + 4 bp and -5 bp for reads mapping to the plus and minus strands, respectively [4] (Buenrostro 2013). Duplications are removed by Picard tools. 7. MACS2 is applied on each merged bam file to call peaks (with option -nomodel -extsize 200 -shift 100). 8. HOMER is used to annotate ATAC-seq peaks to the various genomic regions (promoter, 5 UTR, exon, intron, 3 UTR, downstream, and intergenic), and assign them to the nearest gene based on rat genome assembly rn6. Notably, promoters are defined as 3 kb upstream and downstream from the TSS. Downstream regions correspond to the 3 kb DNA region located downstream from the transcription termination site. 9. All sequencing tracks, bigWig files are viewed using the Integrated Genomic Viewer. Foldchange of ATAC-seq signal is calculated using DiffBind package in R. Interpretation is limited to peaks that exceed a L2FC of 1 (FDR < 0.01) in either direction.

Declaration of Competing Interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.