Quality control and evaluation of plant epigenomics data

Abstract

Epigenomics is the study of molecular signatures associated with discrete regions within genomes, many of which are important for a wide range of nuclear processes. The ability to profile the epigenomic landscape associated with genes, repetitive regions, transposons, transcription, differential expression, cis-regulatory elements, and 3D chromatin interactions has vastly improved our understanding of plant genomes. However, many epigenomic and single-cell genomic assays are challenging to perform in plants, leading to a wide range of data quality issues; thus, the data require rigorous evaluation prior to downstream analyses and interpretation. In this commentary, we provide considerations for the evaluation of plant epigenomics and single-cell genomics data quality with the aim of improving the quality and utility of studies using those data across diverse plant species.

Figure 1

Accounting for mappability and genome assembly artifacts. A schematic diagram of typical aligned data obtained by ATAC-seq. Regions of chromatin accessibility are indicated by peaks. A region that appears enriched for sequencing coverage (labeled “false peak”) is actually due to a collapsed repeat in the genome assembly. Tn5-treated genomic DNA helps to identify these problematic regions. A k-mer-based approach is used to reveal regions of the genome that are uniquely mappable for a given sequence fragment length.

Figure 2

Visualization of ChIP-seq enrichment of histone modifications. Low-quality/failed versus high-quality ChIP-seq data are shown for H3K4me3 and H3K27me3 from soybean (Glycine max) leaves. The first and third tracks show low-quality and/or failed ChIP-seq data, whereas tracks 2 and 4 show high-quality data. Box 1 shows a region of H3K27me3 enrichment in track 4, whereas the same region shows almost no enrichment in track 3. As is typical for H3K27me3, enrichment is present throughout the gene body into the upstream region. Boxes 2 and 3 show enrichment for H3K4me3 at TSSs in track 2, whereas weak enrichment is detected in track 1.