GEO Accession viewer

NCBI > GEO > Accession Display

Not logged in | Login

GEO help: Mouse over screen elements for information.

Sample GSM2111729

Query DataSets for GSM2111729

Status

Public on Dec 18, 2017

Title

srsf2.rep1

Sample type

SRA

Source name

mESC

Organism

Mus musculus

Characteristics

strain: 129/Ola
antibody: srsf2
antibody catalog#: sc53518
antibody vendor: santa cruz

Treatment protocol

MLL3/4 DKO cells were obtained as referenceed in the manuscript.

Growth protocol

mESCs were propagated on feeder cultures then passaged twice feeder free before crosslinking and harvesting.

Extracted molecule

genomic DNA

Extraction protocol

Crosslinkedand sonicated chromatin was incubated with the ab indicated. Complex associated DNA was purified and sequenced.
Libraries were constructed as per manufacturers instructions for Illumina Tru-seq.

Library strategy

ChIP-Seq

Library source

genomic

Library selection

ChIP

Instrument model

Illumina HiSeq 2500

Data processing

Sequencing Read Alignment: ChIP-seq libraries were sequenced on the Hi-Seq 2500 platforms. 36bp single end sequencing data was aligned to a reference mouse genome (mm9) downloaded from UCSC genome browser by using Bowtie. Unmapped and non-uniquely mapped reads were removed, and PCR duplicate reads were removed with Picard.
Peak Identification: H3K4me1 and H3K4me3 peaks were defined by using MACS2 with default parameters except following options “–m 5 50 –p 1e-5” for each biological replicate. After that we selected the best quality replicate for each modification based on number of identified peaks. From the best quality replicate, we only selected reproducible peaks at least two biological replicates. H3K4me1 and H3K4me3 peaks were merged together and classified into H3K4me1 peaks, H3K4me3 peaks, and common H3K4me1/me3 peaks. Chromatin regulator (CR) peaked regions were defined by comparing predefined H3K4me1/me3 peaked regions. Due to the limited quality of CR ChIP-seq results we first calculated input normalized CR ChIP-seq RPKMs on H3K4me1/me3 peaked regions for each biological replicate. If CR ChIP-seq RPKMs is enriched more than 1.5 fold compared to input RPKMs at least two biological replicates the h3k4me1/me3 peaked region was defined as occupied by the corresponding CR. H3K27ac ChIP-seq data was processed similarly as described in CR ChIP-seq data analysis except using 2-fold enrichment threshold when we define H3K27ac peaked regions. The density of histone modifications and CR binding was defined by taking average input normalized RPKMs of reproducible peaks across multiple biological replicates.
Clustering of CR-binding patterns: K-means clustering was performed for CR binding patterns according to poised enhancer and active enhancer regions. The peaks occupied by H3K4me1 but depleted by H3K4me3 were considered as enhancer regions after excluding upstream and downstream 2.5kb from RefSeq transcription start sites. Active enhancers were defined if the enhancer regions were overlapped with H3K27ac peaked regions (n=13,811). Otherwise the region was defined as poised enhancer regions (n=28,008). We performed K-mean clustering based on 13 CR binding patterns at active and poised enhancer regions separately, and generated 15 different clusters. Based on the enrichment of CR binding patterns in each cluster we manually assigned each cluster as ‘No CR bind’, ‘CR-specific bind’, and ‘Multiple CR bind’.
Identification of KMT2C/D dependent and independent H3K4me1 and H3K4me3 peaks: We defined KMT2C/D dependent and independent sites based on H3K4me1 levels between WT and KMT2C/D DKO cell lines. If the region is occupied by H3K4me1 more than 2-fold compared to input data and depleted more than 2-fold in KMT2C/D DKO cell lines the region was defined as a KMT2C/D dependent site, otherwise defined as a KMT2C/D independent site.
Analysis of CR-binding patterns in KMT2C/D DKO cell lines: To determine whether CR binding patterns are dependent on H3K4me1, we performed additional ChIP-seq for BRG1, SMC3, CHD1, PHF5A, MED23, PHRF1, and POUF51 for KMT2C/D DKO cell lines. BRG1, SMC3, PHF5A, MED23, and PHRF1 were selected based on the strong association with H3K4me1 peaked regions in WT. POUF51 was selected as a negative control because POUF51 can be considered to bind as H3K4me1 independently. Sequence read alignment was performed as described in WT CR ChIP-seq data. During downstream analysis of DKO cell lines we only considered CR ChIP-seq peaked regions defined in WT, resulting in 1939 peaks for BRG1, 20,412 peaks for CHD1, 12,225 peaks for POUF51, 2793 peaks for PHF5A, 7,425 peaks for PHRF1, 9,614 peaks for SMC3, and 6,712 peaks for MED23.
Genome_build: mm9

Submission date

Apr 07, 2016

Last update date

May 15, 2019

Contact name

Bing Ren

Organization name

University of California, San Diego School of Medicine

Street address

9500 Gilman Dr., CMM-East, Admin Area/Rm 2071

City

La Jolla