A reassessment of DNA immunoprecipitation-based genomic profiling

Abstract

DNA immunoprecipitation sequencing (DIP-seq) is a common enrichment method for profiling DNA modifications in mammalian genomes. However, DIP-seq profiles often exhibit significant variation between independent studies of the same genome and from profiles obtained by alternative methods. Here we show that these differences are primarily due to intrinsic affinity of IgG for short unmodified DNA repeats. This pervasive experimental error accounts for 50 -99% of regions identified as 'enriched' for DNA modifications in DIP-seq data. Correction of this error profoundly alters DNA modification profiles for numerous cell types, including mouse embryonic stem cells, and subsequently reveals novel associations between DNA modifications, chromatin modifications and biological processes. We propose new methodological guidelines that minimize the impact of these errors on future DIP-seq experiments and allow new insights to be made from the wealth of existing DIP-seq data.

Regulation of transposable elements by DNA modifications

“…With 6mA levels varying widely across species ( Figure 6), sensitive assays have been required to allow the detection of extremely low levels of this modification. The presence of such low 6mA levels has not only been questioned from a technical perspective 34,35 , but also triggered a discussion about the potential functional importance of 6mA in these cases. Nevertheless, a number of different roles have been suggested for 6mA, with TE regulation being a recurrent theme across several species.…”

“…This is an ongoing technical challenge, as a recent ultrasensitive mass spectrometry approach could not confirm the presence of 6mA in mouse ESCs 34 . Genome-wide mapping of 6mA using antibody-based enrichment techniques is also particularly prone to artefacts due to the high background resulting from the affinity of immunoglobulin G (IgG) for simple repeats 35 .…”

“…Apart from 5mC, the most common types of enzyme-catalysed DNA modifications include N4-methylcytosine (4mC) and N6-methyladenine (6mA), which are widespread across bacteria. Notably, 6mA is also found in varying amounts in eukaryotes 30 and has recently been implicated in TE regulation in Drosophila melanogaster and mice 31,32 , although the presence and functional relevance of 6mA in higher eukaryotes remains controversial [33][34][35] . Moreover, 5mC can be oxidized by teneleven translocation (TET) enzymes to 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC) and 5carboxylcytosine (5caC), as part of a replication-independent pathway to DNA demethylation ( Figure 3).…”

Sources of artifact in measurements of 6mA and 4mC abundance in eukaryotic genomic DNA

“…The development of more quantitative sequencing methods, equivalent to bisulfite sequencing for detection of 5mC, are required to thoroughly analyze the dynamics of 6mA during early development. We believe that more mechanistic studies, including the manipulation of potential 6mA methyltransferases and demethylases, are required to thoroughly dissect the potential roles of 6mA at early development and to support the IF and IP results which can be prone to false positive signals [41,42].…”

DNA (Hydroxy)Methylation in T Helper Lymphocytes

“…First, they are not single-nucleotide resolution methods, since they allow the identification of modified regions, but with no information about specific nucleotides. Second, like all antibody-based methods, the utilized antibodies should be carefully controlled for specificity and validated independently, in order to ensure correct interpretations of the data [71]. Techniques that can discriminate 5hmC from other forms of cytosine at a single-base resolution include TET-assisted bisulfite sequencing (TAB-seq) [72] and oxidative bisulfite sequencing (oxBS-seq) [73].…”