ATAC-seq in Emerging Model Organisms: Challenges and Strategies

Abstract

The Assay for Transposase-Accessible Chromatin with sequencing (ATAC-seq) is a versatile and widely utilized method for identifying potential regulatory regions, such as promoters and enhancers, within a genome. ATAC-seq has been successfully applied to a wide range of established and emerging model organisms. However, implementing this method in emerging model systems, such as arthropods, can be challenging due to several factors that influence data quality. These factors include the availability of a sufficient amount and quality of tissue or cells, the need for species- and tissue-specific protocol optimization, the completeness and accuracy of the reference genome, and the quality of the genome annotation. In this article, we emphasize the key steps in the ATAC-seq protocol that, based on our experience, have the greatest impact on data quality when adapting this method for emerging model organisms. Specifically, we discuss the importance of nuclei isolation, the incubation conditions of the Tn5 transposase, and PCR amplification of the library. Furthermore, we outline essential quality checkpoints during the bioinformatic analysis of ATAC-seq data to assist in assessing data integrity and consistency. Given that many emerging model organisms may not be readily available in laboratory cultures, we also emphasize the importance of evaluating how different preservation methods affect ATAC-seq data quality. Based on examples in one spider and one ant species, we demonstrate that replication and thorough quality controls at all steps of the protocol and data analysis are essential to assess the usability of ATAC-seq data. Our data highlights the importance of isolating the right number of intact nuclei, as well as ensuring optimal amplification conditions during library preparation to obtain good-quality sequence data for downstream analyses. We recommend using fresh tissue samples if possible because we show that direct cryopreservation of the tissue may affect chromatin integrity. This effect could be avoided or reduced by preserving the homogenate in cell culture medium. Overall, we explain the ATAC-seq protocol and downstream analyses in detail and give step-by-step advice to researchers who are new to the field and want to implement this method. With careful planning and validation, ATAC-seq can reveal the regulatory landscape of a genome and aid in identifying elements that govern gene expression.

Keywords: ATAC‐seq; benchmarking; chromatin accessibility; gene regulation; insects; protocol optimization; quality control; spiders; tissue preservation.

Figure 1

(A) Phylogeny of species with published ATAC‐seq studies, based on Supporting Information S5: Table S1. The boxes at the tips of the tree indicate the type of sample preparation used (green = fresh material, blue = cryopreserved material, orange = preservation in cell culture medium). Data for the two species labeled with asterisks are generated in this study. The animal silhouettes were downloaded from http://phylopic.org/, with credits to Mathilde Cordellier for the spider Parasteatoda tepidariorum (https://creativecommons.org/licenses/by-nc/3.0/), Thomas Hegna for Artemia salina (https://creativecommons.org/publicdomain/mark/1.0/), T. Michael Keesey for the malacostraca example (https://creativecommons.org/licenses/by-sa/3.0/), Guillaume Dera for Cloeon dipterum, T. Michael Keesey for the ant, Lubna Maherally for the honey bee, Max Farnworth for Tribolium castaneum, Andy Wilson for the Drosophila melanogaster, Oncopeltus fasciatus as example for Hemiptera and Danaus plexippus, Gemma Martínez‐Redondo for Bombyx mori, Mattia Menchetti for Papilio machaon (https://creativecommons.org/publicdomain/zero/1.0/). (B) Overview of the ATAC‐seq method, illustrating the key steps that lead to nucleotide fragments of different size classes. Note that the length of the nucleotide sequencing adapters (shown in green and yellow) is downscaled for clarity. (C) Schematic of example fragment size distributions for ATAC‐seq libraries, as determined by instruments such as Agilent Bioanalyzer or Tapestation. The range of acceptable fragment sizes is indicated by the green‐shaded areas. The purple line shows the fragment size distribution of an over‐tagmented library, while the yellow dotted line illustrates the fragment size distribution of an under‐tagmented library. Red dotted lines indicate the lower and upper fragment markers. NF = nucleosome‐free fraction.

Figure 2

Schematic representation of samples used in this study and the comparison of sample processing methods using either no preservation method (in the following referred to as fresh samples, green box), cryopreservation (in the following referred to as frozen sample, blue box) or cell culture medium based cryopreservation after homogenization (in the following referred to as medium sample, right box). For the ant Temnothorax longispinosus, nurse fat body tissue was used to compare a single fresh and a single cryopreserved sample. For the spider Parasteatoda tepidariorum, all three methods were compared with 3–4 replicates of whole embryos from later embryonic stages during germ band retraction and dorsal closure (Stage 12/13 according to Mittmann and Wolff 2012).

Figure 3

Post‐mapping quality assessment of ATAC‐seq samples. (A) Fragment size distribution of the spider samples using different preservation methods: fresh (green), frozen (blue), medium (orange). Mean values are shown for the replicates. Individual replicate fragment size distributions can be found in Supporting Information S2: Figure S2A. (B) Fragment size distribution of the ant samples for fresh (dark green), fresh subsampled (light green), and frozen (blue) preservation methods. (C) Library complexity estimation for the spider samples using different preservation methods: fresh (green), frozen (blue), medium (orange). Mean values (lines) and standard deviation (transparent shading) are shown for the replicates. Dotted lines show the distinct replicates, and individual replicate fragment size distributions can be found in Supporting Information S2: Figure S2B. (D) Library complexity estimation of the ant samples for fresh (dark green), fresh subsampled (light green), and frozen (blue) preservation method. (E and F) Transcription start site (TSS) enrichment plots for the nucleosome free fraction of the spider replicates (E) and ant (F) samples. Note that the data is only comparable within species and not between species.

Figure 4

Post‐peak calling quality assessment of ATAC‐seq samples. (A) Comparison of lg(‐lg(q‐values)) across treatments for the spider samples. A high lg(‐lg(q‐value)) indicates that a peak is highly statistically significant (an FDR of 0.05 corresponds to 1.301 in the plots) and less likely to be a false positive. The cumulative q‐values for all replicates are used. Refer to Supporting Information S2: Figure S2C for individual plots for each replicate. All pairwise comparisons were significantly different (Pairwise Wilcoxon Rank Sum Tests; Fresh – Frozen: W = 35,503,230,972, p < 2e⁻¹⁶; Fresh – Medium: W = 49,855,901,049, p < 2e⁻¹⁶; Frozen – Medium: W = 43,377,046,641, p < 2e⁻¹⁶). (B) Comparison of lg(‐lg(q‐values)) across treatments for the ant samples. A high lg(‐lg(q‐value)) indicates that a peak is highly statistically significant (an FDR of 0.05 corresponds to 1.301 in the plots) and less likely to be a false positive. All pairwise comparisons were significantly different (Pairwise Wilcoxon Rank Sum Tests; Fresh – Frozen: W = 290,784,030, p < 2.2e⁻¹⁶; Fresh 45% – Frozen: W = 277,107,730, p < 2.2e⁻¹⁶; Fresh – Fresh 45%: W = 261,656,932, p < 2.2e⁻¹⁶). Note that higher values for lg(‐lg(q‐values) represent high confidence in the called peaks. (C and D) Fraction of peaks located in different genomic regions in (C) spider samples and (D) ant samples. (E and F) Venn diagrams depicting the number of shared and unique peaks across treatments in (E) spider and (F) ant samples are shown. The total number of peaks called per treatment is given in parentheses next to the treatment.