 |
 |
GEO help: Mouse over screen elements for information. |
|
| Status |
Public on Nov 20, 2015 |
| Title |
MATa 5C Rep1 |
| Sample type |
SRA |
| |
|
| Source name |
MATa 5C
|
| Organism |
Saccharomyces cerevisiae |
| Characteristics |
genotype: MATa, ho::LYS2, lys2, ura3, nuc1D::LEU2, pGAL::HO::URA3
strain: SK1
media: -ura
tag: NA
|
| Treatment protocol |
NA
|
| Growth protocol |
Yeast from glycerol stocks were streaked out onto agar plates with the appropriate media and incubated at 30°C. Three colonies were picked and grown over night in a 6ml culture of the appropriate media at 30°C while shaking at 250 RPM. The culture was backed down to an OD600 of 0.2 in 100ml of media and grown over night at 30°C with shaking. The culture was diluted to an OD600 of 0.2 in 400ml of media and incubated to mid-log phase at 30˚C with shaking.
|
| Extracted molecule |
genomic DNA |
| Extraction protocol |
Crosslinked cells were lysed by grinding in a cold mortar and pestle
[Hi-C] The yeast culture was crosslinked with 37% formaldehyde to a final concentration of 3% in the culture media for 20 min at 25˚C. The crosslinking was quenched by adding 2X the volume of formaldehyde used of 2.5M Glycine to the culture and incubated culture for 5 min at 25˚C. Cells were lyses by grinding in a cold mortar and pestle. The chromatin was solubilized by adding 1% w/v SDS and Incubating at 65˚C water bath for 10 min. Excess SDS was quenched using 10% v/v Triton X-100. The chromatin was digested with HindIII at 37˚C overnight in a water bath. Digested ends were filled in with biotin-14-dCTP with Klenow at 37˚C for 2 hr in a water bath. The klenow was inactivated by Incubating in a 65˚C water bath for 20 min. Crosslinked and digested fragments were ligated together in a dilute reaction inorder to favor intra-molecular ligation in a 16˚C water bath for 8 hr. Crosslinkes were reversed by incubating the ligation mixture with proteinase K in a 65˚C water bath. Ligation products were purified with phenol (pH 8.0):chloroform and precipitated with sodium acetate pH 5.2 and ethanol. The precipitated DNA was resuspended in 1X TE and then the DNA was re-extracted with phenol (pH 8.0):chloroform and precipitated with sodium acetate pH 5.2 and ethanol. The DNA pellet was dried thoroughly and excess salt was washed out of the solution using an amicon 30kDa spin column using 1X TE. Then the sample was concentrated using a 30kDa amicon column. RNA was degrade using DNase free RNase A at 37˚C for 1 hr. The Hi-C library was quantified on a 0.8% agarose gel in 0.5X TBE by comparing it to known concentration standards. The Hi-C efficiencey was accessed by first PCR amplifing a neighboring interaction in the Hi-C library. This amplicon was split and the aliquots were digested with either HindIII, NheI, or HindIII and NheI at 37˚C overnight. The digested products were quantified on a gel and the percent Hi-C efficiency was calculated as the percent digested with NheI divided by the percent digested in the NheI and HindIII combined reaction. Biotin was removed from un-ligated ends using T4 DNA polymerase at 20˚C for 4hrs and then at 75˚C for 20 min to inactivate the enzyme. The DNA was sheared to 50-700bp using the Covaris S2 sonicator. The DNA ends were repaired using T4 DNA Polymerase, T4 Polynucleotide Kinase, and Klenow fragment of DNA Polymerase I at 20˚C for 30 min. The DNA was purified with one Qiagen MinElute Column. A dATP was added to the 3’ end of the molecules by using Klenow Fragment (exo-) at 37˚C for 30 min and then at 65˚C for 20 min to inactivate the enzyme. The end repaired and A-tailed library was fractionated to 100-300bp using 0.9x and then 1.1x ampure XP extractions. For each 1.0 ng of biotinylated ligation products, 1.0 µl of Dynabeads MyOne Strepavidin C1 beads were used to enrich the library. Illumina PE adapters were ligated to the library using NEB Quick Ligase at room temperature for 15 min. The library was amplified using Illumina primers PE1.0 and PE2.0 for as few cycles as possible. Cycle number was determined by a titration experiment that revealed which cycle would produce enough library for sequencing but didn’t produce higher molecular weight artifacts. The library was purified once more with 1.8x Ampure XP to remove primers and primer dimers. The final library was quantified on a bioanalyzer. The libraries were sequenced on either the GAII or HiSeq platforms.
[5C] The yeast culture was crosslinked with a final volume of 3% formaldehyde (37% w/v, Fisher, Catalog #: BP531-500) in the culture media at 25˚C for 20mins with shaking. The remaining formaldehyde was quenched with an excess of 2.5M glycine. The yeast cells were harvested by centrifugation and lysed by grinding with a mortar and pestle in the presence of liquid nitrogen. The chromatin was digested with HindIII (NEB, R0104S) and the resulting crosslinked restriction fragments were ligated with T4 DNA polymerase (Invitrogen, 15224-090) in a dilute reaction so that intra-molecular ligation was favored. The crosslinking was reversed by incubating overnight at 65˚C in the presence of proteinase K (Invitrogen, 25530-015). The chimeric DNA molecules were purified with a series of phenol:chloroform pH 8.0 (1:1) extractions and precipitated with 100% cold ethanol and 3M sodium acetate pH 5.2. The RNA was degraded with RNase A (Roche, Catalog #: 10109169001) and the yield was quantified using molecular weight standards run on an agarose gel. 3C-PCR was performed on all 3C libraries with 3C PCR primers which represent a variety of genomic distances. Libraries which showed the characteristic, exponential decay of 3C-PCR product abundance relative to genomic distance passed quality control. For one 5C reaction, 4 million genome copies (53.3ng) worth of 3C library were mixed with 1fmol of each of the 5C probes in a 40 µl reaction. The reaction was heated to 95˚C for 9mins to melt the DNA, then slowly cooled to 55˚C to allow for the 5C probes to anneal to the 3C ligation products for a minimum of 4hrs and a maximum of 16hrs. The annealed probes were then ligated together using Taq DNA Ligase (New England Biolabs, M0208S) at 55˚C for 1hr. The 5C products were amplified with a high fidelity polymerase (Amplitaq Gold, Life technologies, 4398813) and 1.6µM of each of the universal primers (Universal_forward: 5’ -/5Phos/CCTCTCTATGGGCAGTCGGTGAT – 3’, Universal_reverse: 5’ – /5Phos/CTGCCCCGGGTTCCTCATTCTCT – 3’). Five 5C reactions for a total of 20 million genome copies were used to produce each 5C library. The Illumina Paired-end sequencing adapters and primers were used to add the Illumina Paired-end sequences to the 5C molecules in order to prepare them for sequencing on either the Illumina GAII or HiSeq platforms.
|
| |
|
| Library strategy |
OTHER |
| Library source |
genomic |
| Library selection |
other |
| Instrument model |
Illumina Genome Analyzer IIx |
| |
|
| Description |
library strategy: 5C
supplemental_table_3_5Cprobes_mapped_and_filtered_to_SK1.xlsx
supplemental_table_4_5Cprobes_mapped_and_filtered_to_w303.xlsx
supplemental_table_S1_all5Cprobes.xlsx
supplemental_readme.txt
significant_differences_1repVs1rep.tar.gz
significant_differences_3repVs3rep.tar.gz
y3512LR_SK1v3-yJB2-R1
|
| Data processing |
[Hi-C, mapping] Iterative mapping and error correction of the chromatin interaction data were performed as previously described (Imakaev, M. et al. Iterative correction of Hi-C data reveals hallmarks of chromosome organization. Nat. Methods 9, 999-1003, doi:10.1038/nmeth.2148 (2012))
[Hi-C normalization] The total read count for each library was normalized to a common value
[Hi-C, comparisons] subtractions and log2 ratios were performed
[5C, mapping] fastq files were mapped to the 5C probe pool using our in-house 5C mapping pipeline. Each side of the paired-end reads was mapped independently to the 5C probe pool using the novoalign mapping software (V2.05 novocraft.com) and 5C interactions were assembled from those paired end reads where both sides mapped uniquely to a single 5C primer of opposite type (forward-reverse or reverse-forward). Since our experiments were carried out using strains which have a different background than that for which the probes were designed to, we lifted the probe coordinates over to the correct genome assembly. For SK1 the 5C probes were aligned to the SK1 genome produced by Scott Keeney’s Group (http://cbio.mskcc.org/public/SK1_MvO/) using the BLAST command line tool blastall (http://0-www-ncbi-nlm-nih-gov.brum.beds.ac.uk/staff/tao/URLAPI/blastall/) using the blastn program and a 20bp word size. This genome assembly contained what appeared to be a 29,701bp translocation from chromosome XI inserted at position 3,623 in the left arm of Chr. III. Our 5C data did not show evidence for a large insertion in Chr. III. Therefore this region of the assembly is likely an assembly error so we manually removed this region from the SK1 genomic sequence (chrIII:3,623-33,324).Once the probes were aligned to the genome they were manually curated to remove those which blasted to the wrong chromosome or to multiple chromosomes. Probes were also filtered if they no longer aligned adjacent to a HindIII restriction site in the SK1 genome. Finally, since the length of the rDNA on the right arm of chromosome XII contains many copies of the rDNA repeats which could, in reality be up to 2Mbps of DNA, we decided to treat the left and right portions of chromosome XII as separate chromosome in our analysis. We did this by changing the chromosome name of these probes to the right of the rDNA to chr12R and restarting the coordinates at 0. The procedure for mapping to the 5C datasets from the strains used for the microscopy experiments (Supplemental Figure S8) was the same as for mapping to the SK1 genome, except that the W303 genome.
[5C, filtering] Prior to correcting the 5C counts using the Iterative Correction algorithm it was necessary to filter the data of any outliers. First, the 70kb around Centromere-centromere and telomere-telomere interactions were removed to prevent from filtering high 5C counts from these regions which are known to interact and thus will have a higher signal than expected. Then single data points which were greater than or equal to 7 standard deviations from the local mean, based on genomic distance, were removed for intra-chromosomal interactions. For inter-chromosomal interactions a cutoff of 8 standard deviations from the mean of the inter-quartile (25%-50%) were removed. Finally, we removed the highest and lowest 7.5% (15% total) of rows and columns from the dataset which were determined by summing all of the inter-chromosomal interactions for each 5C fragment. The centromeres and telomeres were replaced and used for all further correction/analysis. The data was binned into 10kb x 10kb bins. The median of the interactions in each 10kb x 10kb region was used to represent that region. The iterative correction algorithm assumes that all bins should have the same sum of interactions. This assumption is valid when genome wide Hi-C data is being considered. However, the 5C design is comprehensive enough to approximate this assumption. The row and column sums were calculated for each 10kb bin. Then each row and column was multiplied by a factor that would raise or lower its sum to match the average of all the sums. This was done iteratively until the row and column sums deviated very little from each other. Figure S3 shows the sums of each row in the dataset before and after applying the iterative correction algorithm.
[5C, comparisons] In order to call significant differences between pairs of 5C datasets, we tested for the difference in the ranks of interaction frequencies obtained between two pairs of genomic regions. To this end, we first bin the 5C data into 30Kb by 30Kb bins that overlap each other by 10Kb. For each of these bins we pool all 5C interactions that fall within it across the three biological replicates for that strain. We use this list of pooled values for the first strain and perform a two-tailed rank sum test with the list of corresponding values for the second strain. We correct for multiple testing and plot the log2 of the median of strain 1 divided by the median of strain 2 for that bin in the heat map if the corrected p-value from the test between these two bins is less than or equal to a 5% FDR.Filtering, correction, and normalization of 5C data. Prior to correcting the 5C counts using the Iterative Correction algorithm it was necessary to filter the data of any outliers. First, the 70kb around Centromere-centromere and telomere-telomere interactions were removed to prevent from filtering high 5C counts from these regions which are known to interact and thus will have a higher signal than expected. Then single data points which were greater than or equal to 7 standard deviations from the local mean, based on genomic distance, were removed for intra-chromosomal interactions. For inter-chromosomal interactions a cutoff of 8 standard deviations from the mean of the inter-quartile (25%-50%) were removed. Finally, we removed the highest and lowest 7.5% (15% total) of rows and columns from the dataset which were determined by summing all of the inter-chromosomal interactions for each 5C fragment. The centromeres and telomeres were replaced and used for all further correction/analysis. The data was binned into 10kb x 10kb bins. The median of the interactions in each 10kb x 10kb region was used to represent that region. The iterative correction algorithm assumes that all bins should have the same sum of interactions. This assumption is valid when genome wide Hi-C data is being considered. However, the 5C design is comprehensive enough to approximate this assumption. The row and column sums were calculated for each 10kb bin. Then each row and column was multiplied by a factor that would raise or lower its sum to match the average of all the sums. This was done iteratively until the row and column sums deviated very little from each other. Figure S3 shows the sums of each row in the dataset before and after applying the iterative correction algorithm. The 5C data for each dataset was first normalized to a standard read depth. Each interaction in the dataset was calculated as a percentage of the total number of reads for the dataset. This percentage was then multiplied by the arbitrary value of 1,000,000. Finally, 5C data was normalized relative to the expected value for pairs of loci separated by a give genomic distance (as decribed in (Sanyal et al., 2012)). This distance-dependent expected value was calculated using the ârobust lowessâsmoothing method which takes an âoutlier filteredâ weighted average across a percentage of the data (here we used 5%) for each genomic distance. #5C data analysis. In order to call significant differences between pairs of 5C datasets, we tested for the difference in the ranks of interaction frequencies obtained between two pairs of genomic regions. To this end, we first bin the 5C data into 30Kb by 30Kb bins that overlap each other by 10Kb. For each of these bins we pool all 5C interactions that fall within it across the three biological
|
| |
|
| Submission date |
Oct 09, 2015 |
| Last update date |
May 15, 2019 |
| Contact name |
Bryan R Lajoie |
| E-mail(s) |
bryan.lajoie@gmail.com
|
| Organization name |
UMMS
|
| Department |
Program in Systems Biology
|
| Lab |
Dekker Lab
|
| Street address |
368 Plantation St.
|
| City |
Worcester |
| State/province |
MA |
| ZIP/Postal code |
01605 |
| Country |
USA |
| |
|
| Platform ID |
GPL13272 |
| Series (1) |
| GSE73890 |
The conformation of yeast chromosome III is mating type-dependent and controlled by the recombination enhancer |
|
| Relations |
| BioSample |
SAMN04158457 |
| SRA |
SRX1322510 |
| Supplementary file |
Size |
Download |
File type/resource |
| GSM1905029_y3512LR_SK1v3-yJB2-R1-filtered.txt_10000_1.balanced.matrix.gz-n2r.txt.log2ratio.matrix.gz |
83.9 Kb |
(ftp)(http) |
MATRIX |
SRA Run Selector |
| Raw data are available in SRA |
| Processed data provided as supplementary file |
| Processed data are available on Series record |
|
|
|
|
 |
