 |
 |
GEO help: Mouse over screen elements for information. |
|
| Status |
Public on Oct 31, 2018 |
| Title |
WT_cntrl_HiC_rep1 |
| Sample type |
SRA |
| |
|
| Source name |
mESC
|
| Organism |
Mus musculus |
| Characteristics |
strain: CAST/Ei x 129/Sv/Jae
cell type: mESC (Tsix-stop)
genotype: WT
|
| Extracted molecule |
genomic DNA |
| Extraction protocol |
We used the in situ Hi-C method of (Rao et al. Cell 2014) to prepare all libraries, using 5-10 million cells. Importantly, we sequenced 20-40 million reads per library. This is a lower sequencing depth than many published Hi-Cs, however since the megadomains are large and prominent feature of the organization of the Xi, this depth is appropriate for detecting the megadomains efficiently and economically.
|
| |
|
| Library strategy |
Hi-C |
| Library source |
genomic |
| Library selection |
other |
| Instrument model |
Illumina HiSeq 2000 |
| |
|
| Description |
AKa1-14
|
| Data processing |
RNA-seq: RNA-seq reads were aligned to the cas (Xa) and mus (Xi) genomes allele-specifically using a previously published pipeline (Kung et al., Molecular Cell 2015)(GSE58242). Following alignment, gene expression levels for each gene were defined using HOMER. Differential expression and fold changes between conditions were calculated using DESeq2. We plotted the cumulative distributions of fold changes for autosomal and X-linked genes and evaluated the significance of any differences between the distributions of the fold changes using the Kolmogorov-Smirnov (KS) test. To examine allele-specific expression from the Xa and the Xi, we summed together allelic reads across both biological replicates and filtered for genes with at least 12 allelic reads in both wild-type and Dxz4∆/∆:FirreXi∆/+and fpm > 0 in all replicates. We also used RNA-seq done in pure hybrid mus or cas fibroblasts to identify and eliminate genes that have incorrect SNP information. We defined escapee genes in a particular condition as genes where at least 10% of allelic reads came from the Xi in either replicate of that condition. We plotted the distribution of expression levels from the Xi (Xi/(Xi+Xa) read counts) for all genes passing our filtered for each replicate. We evaluated the significance in differences of the mean expression level from the Xi using the Wilcoxon Signed Rank Test with Bonferroni correction for multiple hypothesis testing.
ATAC-seq: Attack seq alignment to mm9 was performed exactly as ChIP-seq alignment was performed in (Minajigi et al., Science 2015) (GSE67516). Peaks were called using macs2 with default parameters. Biallelic peaks were identified as peaks with at least 10 alleleic reads in a sample and an Xi:Xa ratio greater than 1/3. Xi-specific peaks were defined as peaks with at least 10 allelic reads and a Xi:Xa ratio less than 1/3. To test whether Xi-specific peaks in wild-type are “restored” (that is: acquire appreciable accessibility on the Xi) in either the Dxz4 or Firre deletion, we plot the wild-type Xa reads on the x-axis and the deletion Xi reads on the y-axis and identify peaks where the deletion Xi/wild-type Xa ratio is greater than ½ (these are peaks where the deletion accessibility level reaches at least half the wild-type accessibility ratio). We also examine the biallelic peaks and plot the wild-type Xi reads on the x-axis and the deletion Xi reads on the y-axis to determine whether the accessibility on the Xi changes for the peaks that are bi-allelic in wild-type.
Hi-C: Hi-C alignment to mm9 was performed according to the method of (Minajigi et al., Science 2015)(GSE67516) The allele-specific Hi-C reads were filtered for quality and uniqueness with HOMER. Custom scripts were used to convert HOMER tag directories into the format accecpted by Juicebox; contact maps were generated using the Juicer tools ‘pre’ command. All Hi-C contact maps visualized in this study are KR-normalized contact maps generated by Juicebox.
Hi-C^2: Reads of each pair were individually mapped to the mus and cas reference genomes using novoalign and merged into Hi-C summary files and filtered using HOMER as previously described (Minajigi et al., Science 2015). For the chrX:70,370,161-71,832,975 captures, 3-4% of mapped and paired reads fell within the target region (0.05% expected based on size of capture region versus genome) and for the chrX:71,832,976-73,511,687 captures, 1-2% of mapped and paired reads fell within the target region (0.06% expected based on size of capture region versus genome). To avoid computational complexities arising from normalization of sparse, non-enriched regions in the Hi-C contact map, only Hi-C interactions falling within the capture region were analyzed further. For each capture, a custom script was used to pull out the filtered Hi-C interactions falling within the target region from the HOMER tag directories. Hi-C contact maps of the capture regions were then generated from these HOMER tags using the ‘pre’ command of Juicer tools. The resulting Hi-C contact maps in .hic format were visualized and normalized with the ‘Coverage (Sqrt)’ option in Juicebox.
4C: 4C analysis was performed using the previously published pipeline in (Chu et al., NSMB 2017) (GSE69887).
Genome_build: mm9
Supplementary_files_format_and_content: bedGraph files display fragment counts per million reads passing filter for 4C experiments. bigWig files display fpm-normalized ATAC-seq coverage. .hic files can be used to display and generate Hi-C contact maps using the Juicebox toolkit. For bedGraph and bigWig files, "comp" tracks are generated using all unique reads, "mus" tracks are generated using mus-specific reads, "cas" tracks are generated using cas-specific reads and "reps" tracks are generated using non-unique reads (reads with multiple alignments are randomly assigned to one position). genetable*.xlsx files are generated from homer and contain fragment count information for each gene in each experiment.
|
| |
|
| Submission date |
Jul 05, 2018 |
| Last update date |
May 09, 2019 |
| Contact name |
John Edward Froberg |
| E-mail(s) |
jfroberg@fas.harvard.edu
|
| Phone |
8479770174
|
| Organization name |
Harvard University
|
| Department |
Stem Cell & Regenerative Biology
|
| Lab |
Macklis Lab
|
| Street address |
7 Divinity Avenue
|
| City |
Cambridge |
| State/province |
MA |
| ZIP/Postal code |
02138 |
| Country |
USA |
| |
|
| Platform ID |
GPL13112 |
| Series (1) |
| GSE116649 |
Megadomains and superloops form dynamically during X-chromosome inactivation but are dispensable for silencing and escape |
|
| Relations |
| BioSample |
SAMN09604737 |
| SRA |
SRX4342898 |
| Supplementary file |
Size |
Download |
File type/resource |
| GSM3244840_AKa1_14.cas.hicsummary.txt_TD_noSDbg1000.chrX.tags.hic |
3.3 Mb |
(ftp)(http) |
HIC |
| GSM3244840_AKa1_14.comp.hicsummary.txt_TD_noSDbg1000.chrX.tags.hic |
4.8 Mb |
(ftp)(http) |
HIC |
| GSM3244840_AKa1_14.mus.hicsummary.txt_TD_noSDbg1000.chrX.tags.hic |
3.4 Mb |
(ftp)(http) |
HIC |
SRA Run Selector |
| Raw data are available in SRA |
| Processed data provided as supplementary file |
|
|
|
|
 |
