1
H2A.Z demarcates intergenic regions of the plasmodium falciparum epigenome that are dynamically marked by H3K9ac and H3K4me3.
PLoS Pathog. 2010; 6(12):e1001223
All selected articles haveThis article has been added to your saved articles
You can follow/unfollow articles via MyF1000 or via the article or browse pages.
07 Jun 2011
New Finding, Technical Advance
DOI: 10.3410/f.11038956.11988054
While investigating the distribution of the H2A.Z histone variant in the highly AT-rich genome of Plasmodium, the authors uncovered an insidious AT bias in next generation sequencing that had resulted in serious misinterpretations of chromatin profiling data, an issue that is likely to be of general concern...Continue reading
While investigating the distribution of the H2A.Z histone variant in the highly AT-rich genome of Plasmodium, the authors uncovered an insidious AT bias in next generation sequencing that had resulted in serious misinterpretations of chromatin profiling data, an issue that is likely to be of general concern in studies based on epigenomic profiling. The authors provided a clever technological solution to the problem and used it to establish that H2A.Z is a constitutive mark of gene regulatory regions in Plasmodium.
Nucleosome occupancy is a key feature of eukaryotic genomic landscape, and detecting variations can provide important insights into gene regulation. For example, a surprising finding reported in several papers in recent years is that exons are more highly occupied by nucleosomes than introns, which led to the intriguing possibility that this difference is involved in regulation of transcriptional elongation. Although it was immediately evident that relative GC richness could account for many differences in nucleosome occupancy, it was important to establish that any such differences had a biological origin and were not the result of an unappreciated technical artifact. For microarray-based studies, this was an obvious concern because hybridization efficiency at any temperature depends strongly on the G+C content of a sequence, but, because most recent studies documenting differences in nucleosome occupancy used deep sequencing, any nucleosome occupancy differences were generally assumed to reflect the in vivo situation. However, compositional biases do indeed occur in short read sequencing, and this issue has special importance in tag-based studies, such as chromatin profiling where the signal is simply the number of tags per sequence interval. It is often difficult to adequately control for this kind of artifact with a naked DNA control because fragmentation of chromatin and of naked DNA can be very different. Thus, it has not been clear to what extent the results of sequence-based chromatin profiling have been impacted upon by technical artifacts.
In this study, the authors found that Illumina short read sequencing of naked Plasmodium DNA produced profiles that essentially matched those reported by others as differences in nucleosome occupancy, where the supposed depletion of nucleosomes from intergenic regions is instead attributable to the near absence of sequence reads from strongly AT-rich DNA. The authors traced the artifact to polymerase chain reaction (PCR) amplification by showing that Illumina sequencing without PCR gave much more even recovery of sequences. But some amplification is needed for typical samples, which are available in limited amounts, and so the authors devised a general linear amplification method for sequence library preparation: an adapter that includes the promoter sequence for T7 RNA polymerase is used such that RNA synthesis followed by reverse transcription yielded sufficiently abundant levels of unbiased double-stranded copies that small amounts of starting DNA could be used. Insofar as the authors detected depletion of even mildly AT-rich sequences, this method or others aimed at reducing PCR-dependent biases {1,2} need to be widely adopted. Note that compositional biases remain even after eliminating biases in library preparation so that caution is warranted in interpreting differences in nucleosome occupancy based on deep sequencing technologies, especially when the differences are subtle and not verified by alternative methods.
Nucleosome occupancy is a key feature of eukaryotic genomic landscape, and detecting variations can provide important insights into gene regulation. For example, a surprising finding reported in several papers in recent years is that exons are more highly occupied by nucleosomes than introns, which led to the intriguing possibility that this difference is involved in regulation of transcriptional elongation. Although it was immediately evident that relative GC richness could account for many differences in nucleosome occupancy, it was important to establish that any such differences had a biological origin and were not the result of an unappreciated technical artifact. For microarray-based studies, this was an obvious concern because hybridization efficiency at any temperature depends strongly on the G+C content of a sequence, but, because most recent studies documenting differences in nucleosome occupancy used deep sequencing, any nucleosome occupancy differences were generally assumed to reflect the in vivo situation. However, compositional biases do indeed occur in short read sequencing, and this issue has special importance in tag-based studies, such as chromatin profiling where the signal is simply the number of tags per sequence interval. It is often difficult to adequately control for this kind of artifact with a naked DNA control because fragmentation of chromatin and of naked DNA can be very different. Thus, it has not been clear to what extent the results of sequence-based chromatin profiling have been impacted upon by technical artifacts.
In this study, the authors found that Illumina short read sequencing of naked Plasmodium DNA produced profiles that essentially matched those reported by others as differences in nucleosome occupancy, where the supposed depletion of nucleosomes from intergenic regions is instead attributable to the near absence of sequence reads from strongly AT-rich DNA. The authors traced the artifact to polymerase chain reaction (PCR) amplification by showing that Illumina sequencing without PCR gave much more even recovery of sequences. But some amplification is needed for typical samples, which are available in limited amounts, and so the authors devised a general linear amplification method for sequence library preparation: an adapter that includes the promoter sequence for T7 RNA polymerase is used such that RNA synthesis followed by reverse transcription yielded sufficiently abundant levels of unbiased double-stranded copies that small amounts of starting DNA could be used. Insofar as the authors detected depletion of even mildly AT-rich sequences, this method or others aimed at reducing PCR-dependent biases {1,2} need to be widely adopted. Note that compositional biases remain even after eliminating biases in library preparation so that caution is warranted in interpreting differences in nucleosome occupancy based on deep sequencing technologies, especially when the differences are subtle and not verified by alternative methods.
References
Disclosures
None declared
Add Comment
Henikoff S: F1000Prime Recommendation of [Bártfai R et al., PLoS Pathog 2010, 6(12):e1001223]. In F1000Prime, 07 Jun 2011; DOI: 10.3410/f.11038956.11988054. F1000Prime.com/11038956#eval11988054
F1000Prime Recommendations, Dissents and Comments for [Bártfai R et al., PLoS Pathog 2010, 6(12):e1001223]. In F1000Prime, 20 May 2013; F1000Prime.com/11038956
Only registered/signed-in users can make comments.
Sign in or Subscribe to create your user account and join the discussion.
Sign in or Subscribe to create your user account and join the discussion.
No comments yet.
Epigenetic regulatory mechanisms and their enzymes are promising targets for malaria therapeutic intervention; however, the epigenetic component of gene expression in P. falciparum is poorly understood. Dynamic or stable association of epigenetic marks with genomic features provides important clues about their function and helps to understand how histone variants/modifications are used for indexing the Plasmodium epigenome. We describe a novel, linear amplification method for next-generation sequencing (NGS) that allows unbiased analysis of the extremely AT-rich Plasmodium genome. We used this method for high resolution, genome-wide analysis of a histone H2A variant, H2A.Z and two histone H3 marks throughout parasite intraerythrocytic development. Unlike in other organisms, H2A.Z is a constant, ubiquitous feature of euchromatic intergenic regions throughout the intraerythrocytic cycle. The almost perfect colocalisation of H2A.Z with H3K9ac and H3K4me3 suggests that these marks are preferentially deposited on H2A.Z-containing nucleosomes. By performing RNA-seq on 8 time-points, we show that acetylation of H3K9 at promoter regions correlates very well with the transcriptional status whereas H3K4me3 appears to have stage-specific regulation, being low at early stages, peaking at trophozoite stage, but does not closely follow changes in gene expression. Our improved NGS library preparation procedure provides a foundation to exploit the malaria epigenome in detail. Furthermore, our findings place H2A.Z at the cradle of P. falciparum epigenetic regulation by stably defining intergenic regions and providing a platform for dynamic assembly of epigenetic and other transcription related complexes.
DOI: 10.1371/journal.ppat.1001223
PMID: 21187892
