normR: Regime enrichment calling for ChIP-seq data

Helmuth, Johannes; N, Li; Arrigoni, Laura; Gianmoena, Kathrin; Cadenas, Cristina; Gasparoni, Gilles; Sinha, Anupam; Rosenstiel, Philip; Walter, Jörn; Hengstler, Jan G.; Chung, Ho‐Ryun

doi:10.1101/082263

Cited by 11 publications

(9 citation statements)

References 74 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We tested MACS, which is frequently used for the analysis of ChIP-Seq data (Zhang et al 2008;Feng et al 2012). Because of the large size of heterochromatin regions like NADs and LADs compared to most transcription factor binding sites, we also tested EDD (Lund et al 2014), hidden domains (Starmer and Magnuson 2016), and normr (Kinkley et al 2016;Helmuth et al 2016), packages suited for detecting enrichments over large genomic regions. We tested each of these using at least two different settings ( Fig.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Two Contrasting Classes of Nucleolus-Associated Domains in Mouse Fibroblast Heterochromatin

Vertii

et al. 2018

Preprint

View full text Add to dashboard Cite

In interphase eukaryotic cells, almost all heterochromatin is located adjacent to the nucleolus or to the nuclear lamina, thus defining Nucleolus-Associated Domains (NADs) and Lamina-Associated Domains (LADs), respectively. Here, we determined the first genome-scale map of murine NADs in mouse embryonic fibroblasts (MEFs) via deep sequencing of chromatin associated with purified nucleoli. We developed a Bioconductor package called NADfinder and demonstrated that it identifies NADs more accurately than other peak-calling tools, due to its critical feature of chromosome-level local baseline correction. We detected two distinct classes of NADs. Type I NADs associate frequently with both the nucleolar periphery and with the nuclear lamina, and generally display characteristics of constitutive heterochromatin, including late DNA replication, enrichment of H3K9me3 and little gene expression. In contrast, Type II NADs associate with nucleoli but do not overlap with LADs. Type II NADs tend to replicate earlier, display greater gene expression, and are more often enriched in H3K27me3 than Type I NADs. The nucleolar associations of both classes of NADs were confirmed via DNA-FISH, which also detected Type I but not Type II probes enriched at the nuclear lamina. Interestingly, Type II NADs are enriched in distinct gene classes, notably factors important for differentiation and development. In keeping with this, we observed that a Type II NAD is developmentally regulated, present in MEFs but not in undifferentiated embryonic stem (ES) cells. microscopy images (Supplemental Fig. 1B; see also the "NAD-seq" QC files at the NIH 4DN Data Portal (data.4dnucleome.org)). Quantitative PCR suggested that rDNA sequences were enriched relative to the input total genomic DNA > 20-fold in the non-crosslinked preparations, and > 7-fold in crosslinked preparations ( Supplemental Fig. 1C). Immunoblot analyses showed enrichment of nucleolar protein fibrillarin relative to non-nucleolar proteins such as actin, a nucleoporin (Nup62), or lamin A/C (Supplemental Fig.1D). Additionally, we characterized the small RNAs present in the non-crosslinked preparation, observing that the nucleoli were highly enriched for the nucleolar U3 RNA (Supplemental Fig.1E). Small RNAs could not be recovered from crosslinked samples (data not shown).Bioinformatic analysis of nucleolar-associated DNA. We performed three biological replicate experiments for both the crosslinked and non-crosslinked protocols. For each experiment, we purified DNA from the isolated nucleoli, alongside genomic DNA from a sample of whole, unfractionated cells from the same population for normalization of the nucleolar sequencing data. These DNAs were used to generate sequencing libraries via PCR-free protocols, yielding genome-scale maps. Visual inspection of these "NAD-seq" data showed that the total genomic DNA was largely evenly distributed across the genome (Fig. 1B, C), occasionally interspersed with short, poorly represented regions. The nucleolar DNA was distributed with distinct pea...

show abstract

Section: Resultsmentioning

confidence: 99%

“…For Hidden Domains, we compared binning at 5, 10, and 50 Kb (Starmer and Magnuson 2016). For normr, we compared binning at 10 and 50 Kb (Kinkley et al 2016;Helmuth et al 2016). As observed for chr5 (Fig.…”

Section: Supplemental Figure S3 Additional Comparisons Of Different mentioning

confidence: 98%

Two Contrasting Classes of Nucleolus-Associated Domains in Mouse Fibroblast Heterochromatin

Vertii

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…Read counts of each feature were normalized to the control with the R package normR (Supplemental Fig. S17; Helmuth et al 2016;Kinkley et al 2016). Although for most features the signal is extracted in a genomic region around the active gene TSS, for broader features such as elongation marks H3K36me3, RNAPII S2P, and H3K79me2, the signal was averaged over the entire gene body (Supplemental Table S7).…”

Section: Definition Of Model Featuresmentioning

confidence: 99%

Kinetics of Xist-induced gene silencing can be predicted from combinations of epigenetic and genomic features

et al. 2019

View full text Add to dashboard Cite

show abstract

“…To assess the severity of the problem genome-wide, we thought of using histone mark ChIP-Seq data in H9 cells from ENCODE, and cluster the 31,912 unique TSS event into groups of similar epigenetic states 15 . To this end, the enrichment of H3K4me3 (ENCFF161EGP ENCODE id), H3K4me1 (ENCFF804YUX), H3K9me3 (ENCFF049TFL), H3K27me3 (ENCFF022SFF), H3K36me3 (ENCFF760NOJ) and H3K27ac (ENCFF271XBU) versus Input (ENCFF734TEC) in the +/−1 kb region surrounding CAGE tag-cluster representatives has been calculated with normR 26 . These enrichment results facilitated the k-means clustering of CAGE tag-clusters in 9 groups, following silhouette coefficient analysis (data not shown).…”

Section: Overview Of Adapt-cagementioning

confidence: 99%

Solving the transcription start site identification problem with ADAPT-CAGE: a Machine Learning algorithm for the analysis of CAGE data

Γεωργακίλας

Perdikopanis

Hatzigeorgiou

2020

Sci Rep

View full text Add to dashboard Cite

Cap Analysis of Gene Expression (CAGE) has emerged as a powerful experimental technique for assisting in the identification of transcription start sites (TSSs). There is strong evidence that CAGE also identifies capping sites along various other locations of transcribed loci such as splicing byproducts, alternative isoforms and capped molecules overlapping introns and exons. We present ADAPT-CAGE, a Machine Learning framework which is trained to distinguish between CAGE signal derived from TSSs and transcriptional noise. ADAPT-CAGE provides highly accurate experimentally derived TSSs on a genomewide scale. It has been specifically designed for flexibility and ease-of-use by only requiring aligned CAGE data and the underlying genomic sequence. When compared to existing algorithms, ADAPT-CAGE exhibits improved performance on every benchmark that we designed based on both annotationand experimentally-driven strategies. This performance boost brings ADAPT-CAGE in the spotlight as a computational framework that is able to assist in the refinement of gene regulatory networks, the incorporation of accurate information of gene expression regulators and alternative promoter usage in both physiological and pathological conditions. Cap Analysis of Gene Expression (CAGE) was initially introduced in 2003 1 as a novel method specifically developed to capture and quantify 5′ ends of capped RNAs. During the last decade, CAGE has been continuously refined and improved into its current mature form as a well-established protocol for the identification of transcription start sites (TSS) and promoter regions of transcribed loci. The FANTOM Consortium 2 has extensively applied CAGE on hundreds of tissues and cell lines to produce a high-quality annotation of the human and mouse promoterome and characterize regulatory mechanisms of gene expression. Despite its increasing popularity as an experimental promoter identification protocol, the specificity of CAGE regarding the identification of transcription initiation events in the genome has several limitations. There is strong evidence 3-5 that besides promoter regions, CAGE also identifies capping sites along various locations of transcribed loci such as different splicing products, isoforms and capped molecules that can be summed up as transcriptional noise. As a result, only a portion of regions enriched in CAGE signal were found to overlap with the surrounding region of annotated TSSs. This poses a significant hindrance to research studies that aim to integrate regulatory regions into the framework of biological pathways. During the last decade, several in silico methodologies have been developed to provide basic pipelines for analyzing CAGE datasets and to facilitate peak identification and annotation. PARACLU 6 performs clustering of CAGE tags into wider regions based on a density parameter that reflects the trade-off between size and number of overlapping reads. RECLU 7 is an adaptation of PARACLU algorithm that is able to handle replicated (maximum two replicates) experiments base...

show abstract

normR: Regime enrichment calling for ChIP-seq data

Cited by 11 publications

References 74 publications

Two Contrasting Classes of Nucleolus-Associated Domains in Mouse Fibroblast Heterochromatin

Two Contrasting Classes of Nucleolus-Associated Domains in Mouse Fibroblast Heterochromatin

Kinetics of Xist-induced gene silencing can be predicted from combinations of epigenetic and genomic features

Solving the transcription start site identification problem with ADAPT-CAGE: a Machine Learning algorithm for the analysis of CAGE data

Contact Info

Product

Resources

About