Please enable JavaScript.
Coggle requires JavaScript to display documents.
probablistic establishment of speckle-associated inter-chromosomal…
probablistic establishment of speckle-associated inter-chromosomal interactions
abstract
intro
known
inter-chromosomal interactions: genome org에 중요 역할
unknwown
organizational principles
here
in situ Hi-C 결과로 inter-chromosomal interactions을 systematically characterizing하는 computational method
methods
validation
DNA Oligopaint FISH (fluorescence in situ hybridization)
inter-chromosomal contacts measure
Hi-C & Oligopaint DNA FISh
results
identified 2 hub-like inter-chromosomal contacts (assoc. w/)
nuclear speckles
: highly cell-type invariant
enrichment of CSEs (cell-type common super-enhancers)
nucleoli
validation
strong but probabilistic interaction btn speckles & CSE-harboring genomic regions
likelihood of speckle-CSE associations -> predict measured inter-chromosomal contacts
pop lv
cumulative effect of summing individual stochastic chromatin-speckle interactions
CSE: highly co-occupied by MAZ binding
MAZ depletion -> disorganization of speckle-associated inter-chromosomal contacts
conclusion
simple organizational principle of inter-chromosmal interactions <- MAZ-occupied CSEs
introduction
inter-chromosmoal interaction approaches
substantial part of genome organization
presence of organizational principles that constrain the overall spatial arrangement of inter-chromosomal interactions
experimental
sequencing
SPRITE, sc-SPRITE, RD-SPRITE
identified: inter-chromosomal hubs around nuclear bodies (speckles, nucleoli)
TSA-Seq
map 3D genome org ~ specific nuclear compartments
2 types of trascription hot zones (dist. speckles)
imaging
genome-scale chromatin imaging
DNA-MERFISh, seqFISh+, OligoFISSEQ, in situ genome sequencing
computational
SNIPER, SPIN
analysis of the genome compartments from Hi-C data
governing principle underlying inter-chromosomal interactions: poorly understood
whether & how inter-chromosomal interaction hubs: organized in cell-type-specific/ independent manners
detailed molecular mechanism
associated genomic elements
tech difficulties -> only applied to a limited # cell types
systematic investigation (multiple cell types, effect on gene regulation): not performed
new method: extraction of Hi-C datasets -> inter-chromosomal interactions
widely available for numerous cell/tissue types
useful to probe the principle
inter-chromosomal interactions
Hi-C
done
genome-wide observation of chromatin interactions
principles & functions of multi-layered genome organization
intra-chromosomal
chromomsal compartmentalization @ multiple length scales
cell-type specific higher-order str => chromatin: well-organized hierarchical str - genome functions
limitation
poor detection efficiency & limited spatial distances (<- proximity ligation)
inter-chromosmal interactions X
here
sparse Hi-C inter-chromosomal contact maps -> low D
non-negative matrix factorization (NMF)
non-negativity property of Hi-C contact maps
identified 2 types of inter-chrosomoal interactions
nuclear speckle
highly conserved across (human) cell types
genomic regions: enriched w/ CSEs (common super-enhancers)
nucleolus
DNA Oligopaint FISH anlayses
probailistic interaction of speckle-associated inter-chromosomal interactions @ sc lv
vs pred: fixed hub-like str
DNA-binding motif analysis
MYC Associated Zinc Finger Protein (MAZ): highly co-occupied with CSEs
-> speckle-associated inter-chromosomal interactions
DNA Oligoplant FISH & Hi-C exp
on MAZ KD cells
inter-chromosomal interactions: probabilisitcally established <- stochastic association btn MAZ-occupied CSEs & nuclear speckles
materials & methods
Cell culture
DNA oligopaint FISH
oligonucleotide probe design
primary probe synthesis
combination of immunofluorescence and primary probe hybridization
instrumental setup
sequential hybridization of readout probes and image aquisition
40 imaging target loci selection
Image analysis
nucleus and speckle localization
drift correction and DNA loci localization
RNA interference
RNA interference
RT-qPCR
Western blot
Chromatin immunoprecipitation sequencing
ChIP-seq library
ChIP-seq analysis
In situ hi-C
RNA sequencing
RNA-seq library
RNA-seq analysis
HiCAN
Hi-C data alignment and normalizaiton
Gene annotation
HiCAN
GB176 HiCAN analysis
HiCAN analysis on MAZ KD/KO hi-C data
TAD boundary
Comparison of HiCAN results with other methods
MERFISH
SPRITE
subcompartment annotation
Super-enhancer analysis
Public histone ChIP-seq data
Super-enhancer annotations
Super-ehancer call
CSE protein motif analysis
Protein sequence analysis
Validation of the probabilistic establishment model
Model-predicted inter-chromosomal contact map
Empirically-measured inter-chromosomal contact map
Quantification and statistical analysis
Image averaging analysis
Spatial distance between genomic locus and speckles
Bootstrapping analysis
Statistical tests
results
HiCAN captures two types of nuclear hub-associated inter-chromosomal interactions
recapitulates previously reported hub-like org? (hubs surrounded by nuclear speckles & nuecleoli
S- & N- bases values: MERFISH data에 기반한 각 locus와 nuclear bodies의 contact frequency 잘 resemble
strong corr.
gene-rich, S-basis values & loci-to-speckle contact freq (1C)
gene-poor, N-basis & loci-to-nucleolus contact frequency (1D)
U-basis: no meaningful relationships with both (S1J, K)
SPRITE
S-basis: enrichment of speckle hub genomic regions (1E)
N-basis: strong assoc. w/ nucleolar hub genomic regions
high S-basis regions (<- HiCAN): nuclear speckles와 가까이 위치해 있는지?
DNA Oligopaint FISH in IMR90
S-basis value 높은 genomic loci (vs random loci)와 nuclear speckles의 close association (sc res)
1G-I
centromere-proximal regions preferentially localize around nucleoli
high values of gene-poor basis (S1L)
publicly available Hi-C results (8 cell types, 500kb res.(?))
2 mjr bases (S, N): consistently extracted
every cell type (S1B, C)
highly reproducible
biological replicates, diff resolutions, diff seq depths (S1D-F)
compartmentalized inter-chromosomal organizations
같은 basis에 강하게 속하는 genomic loci: higher inter-chromosomal contacts (다른 bases에 속하는 애들에 비해) (1B, S1G)
no simple 1-1 relationship btn 2 mjr bases & compartment A/B (S1I)
altho S-basis: enriched in gene density & others: depleted ~ compartment A/B patterns (S1H)
HiCAN captures unique 3D inter-chromosomal org from Hi-C contact maps
HiCAN: extract 3 types of genomic loci exhibiting distinct modes of inter-chromosomal interactions & their association tendency w/ nucear bodies
tight assoc: nucelar bodies & gene-rich/poor bases
gene-rich S-basis: speckle-associated chromosome org
gene-poor N-basis: nucleoulus-associated chromosome org
U-basis: regions devoid of inter-chromsomal interactions
from in situ Hi-C contact maps
Hi-C contact maps에서 inter-chromosomal interactions를 구성하는 genomic regions를 identify하기 위해
Hi-C inter-chromosomal Contact map Anlysis with NMF
core design principle
intra-chromosomal interaction-filled inter-chromosmal Hi-C
contact map
구성
NMF
사용해 contact map을
3d space
에 projection
gene density에 기반해 S (speckle-associated)-, N (nucleolus-associated)-, U (undefined)- basis
annotation
distinct conformations를 extract
Hi-C의 제한 (상대적으로 드문 inter-chromsomal interactions를 감지하는) 극복 ->
numeric value of each entry (genomic locus) in an NMF basis
그 locus가 해당 basis에 얼마나 belong하는지
non-orthogonality & non-negativity of bases (????)
effective deconvolution of the complex intermingled inter-chromosmal interactions into biological interpretable low-D strs. <-
results (S1A)
S- & N- : highly skewed distr. to specific genomic regions exclusive to each other
distinct modes of inter-chromosomal organizations
U-: normal distribution across entire genome
genome-wide bg signals
Cell-type invariant speckle-associated inter-chromosomal interactions
values in S-basis & N-basis depend on cell types?
S: conserved
TSA-seq results =
sequence-encoded nature? ->
cell-type conserved spatial genome territories
(= cell-type-independent genomic elements that drives preferential association with nuclear speckles)
N-/U-: less conservative (2A, B)
seq-specific speckle-associated inter-chromosomal interaction -> chromosomal 탈선: loci-speckle assoc에 큰 영향 x 줄것
cancer cell lines: derivative chromosomes & chromsome aneuploidy
multiple reciprocal translocations
S-basis values of loci near translocation breakpoints: still highly preserved (2C, S2A-C)
translocation 없는 다른 cell types과 비슷한 경향
DNA Oligopaint FISh on HeLa w/ 70-90 chromosomes & > 20 translocations
loci: assoc. w/ nuclear speckles (2D, E)
2 loci (high S-basis values in IMR90 (1G-I))
interactions btn IMR90 & HeLa: conservation (2F)
speckle-associated inter-chromosomal interactions: highly conserved across cell types, <- cell-type invariant genomic elements
CSEs are highly enriched in speckle-associated inter-chromosomal interacting regions
identify speckle-associated inter-chromsomal interactions를 구성하는데 책임이 있는 chromatin marks & regulatory elemnts
4 well characterized cell lines
~prev
active histone marks (H3K4me3, H3K27ac): (+) corr. w/ S-basis values (not w/ N-basis values) (3A, S3A)
unexpected
facultative heterochromatin (H3K27me3): also slight enrichment on S-basis
constitutive heterochroamatin (H3K9me3): generally depleted (3A)
~enrichment test
subnatial portion of subcompartment B1: high S-basis (S1I)
speckle-associated interactions: not only 'active hub', also include some portion of facultative heterochromatin ('inactive hub')
degree of enrichment of super-enhacners
genomic looping, phase separation에 중요 역할 (?)
mainly involved in cohesin-indep. inter-chrom. interactions (?)
dir. related to the speckle-assoc. inter-chrom. interactions
CSE (common super-enhancer)s
substantial portion of SEs
constitutively active across multiple cell types
assoc. w/ rapid chromatin loop recovery (?)
hyp: CSEs -> speckle-assoc. inter-chrom. interactions
cuz S-basis: cell-type-invariant
result: (3A-C)
CSEs: over-represented at top-ranked genomic regions in the S-basis (all 4 lines)
ISEs (intermediate) & SSEs (specific): weakly enriched / depleted
S-basis: CSEs와의 genomic distance에 따라 rapidly decrease
CSEs: anchor speckle-associated inter-chrom. interactions of the nearby genome <- (?)
CSE enrichment: indep. of H3K27ac lv (S3C)
CSEs - high S-basis values regions: NOT <- strength of SE activities (?)
relatively higher res.(100kb)
consistent (S3D)
raw inter-chrom. Hi-C reads: also enriched on the CSE regions vs active genes (S3E, F) (?)
CSE itself: key genomic element -> speckle-assoc. inter-chrom. interactions
DNA oligopaint FISh validates CSE-mediated cell-type invariant speckle-associated inter-chromosomal interactions
측정된 S contact freq -> threshold distance (physical contacts 정의하는데 쓰임)
method
image averaging analysis for each type of loci
result
both cell type: S signals - highly enriched around the CSE loci (not for nCSE loci) (4G, S4B)
radial distribution profiles (?) -> peak of S signals ~ center of avg CSE loci (not nCSE loci) (S4C)
cumulative distr. of dist. to the nearest speckle -> CSE loci: closer to Ss (vs nCSE/ active gene loci far from CSE) (4H, S4D)
CSEs와 nuclear speckles의 association을 sc res에서 validatea하기 위해
result
high assoc. of CSEs (vs nCSEs) to Ss (4B, C)
S contact freq of each locus - strength of inter-chm interactions (S-basis): linear relationship (4D)
HUVEC: similar results (4E, F, S4A)
CSEs - Ss: cell-type-indep. assoc.
S assoc. rates: v. similar btn CSE & CSE-proximal gene pairs (4D, F) (?뭔의민데)
method
DNA Oligopaint FISH -> 40 loci
probes -> 20 CSEs, 10 nCSEs, 2 active/inactive genes close to the CSEs, 2 active genes far from the CSEs, 4 high N-basis regions (4A) (?)
combined immunofluorescnece & DNA Oligopaint FISH -> locations of nuclear speckles & genomic loci (4B)
HiCAN: well predicts both S- & N- assoc. interactions
expr lv of gene itself -> S assoc of the locus?
method
6 genes
result
strong corr: genomic distances from respective CSEs & S assoc.freq
expr lv: NO corr w/ S assoc. (4I)
high N-basis values -> strong assoc. of genomic region w/ N?
high N-basis regions: assoc. w/ N (vs CSEs (?)) (S4E, F)
assoc. rates of indiv CSE loci w/ S: highly conserved (4J, S4G)
S contact freq of each targeted locus: strongly corr. btn 2 cell types
Association probabilities between each CSE and speckles determine inter-chromosmal contact frequencies
model
simple mathematical interpretation of HiCAN's matrix decomposition procedure ->
Hi-C contact matrix M (n x n)
NMF with k = 3 -> matrix W (n x 3), H (3 x n)
each column of W: S/N/U-basis
symmetry of contact matrix -> H = W^T
S-basis value of each genomic locus (Si): Ss에 대한 각 locus i의 contact probability (pi) 를 represent
두 다른 loci i, j가 동시에 같은 S에 contact하고 있을 확률: pi * pj에 비례 <- (?)
각 loci와 Ss의 association rates -> Hi-C에서 측정한 speckle-associated inter-chm contact freq 예측 가능
low number of multi-contacts observed using DNA Oligopaint FISH
각 포유류 세포: 보통 몇십게 Ss 가짐 -> 한 S에 여러 CSE loci가 동시에 co-loc되는 multi-contact events의 probability 줄임
genomic loci와 Ss의 물리적 associations: highly dynamic ->
반면 세포 population에서 한 Hi-C assays - (accumulation of individual inter-chm contacts (5A)) -> hub-like profiles 나타냄
model capability
각 loci-S pair의 co-loc probabilities <- DNA Oligopaint FISH assay에서 얻음 각 loci의 S association rates
probabilistic establishment model -> inter-chm contact map 예측 -> S association rates 측정 (5B)
예측 matrix vs 실험적으로 측정한 contact maps &
Hi-C inter-chm contact map (5D)
DNA-Oligopaint-FISH-derived, 각 loci pair에 대해 같은 S에서의 co-loc events 횟수 셈 (5C)
results
예측된 contact map: resembalance to both maps (5E-G)
various distance thresholds, different cell line -> consistent
model describes the formation of S-associated inter-chm interactions with higher accuarcy
gene transcription scores (sum of FPKM values @ 30kb res.)나 chm activities (H3K27ac peaks) 썼을 때는 그런 정학환 예측 결과 나오지 않았음 (S5E-G)
deterministic model (fixed contact partners for each S) -> poorly resembles the hub-like str of the Hi-C contact map (S5H-J)
S-associated inter-chm interactions: probabilistically shaped <- indep. assoc. of indiv. loci-Ss -> dynamic interactions w/i identical S compartments
SRPITE와 Hi-C 실험에서 관찰된 hub-like str <- 다른 loci가 동시에 같은 S를 contact하는 liklihood를 reflect하는 걸수도
CSEs와 Ss간에 stochastic association behaviors (sc lv)가 어떻게 hub-like inter-chromosomal organization (pop lv) 을 설명할 수 있는지 묘사
stochastic assoc.
DNA Oligopaint FISH에선 여러 CSE loci가 같은 speckle에 co-localize되는 multi-contact events 관찰 거의 못함
Hi-C랑 SPRITE에는 hub-like chromatin interactions 있었음 (1B)
cell마다 같은 S에 co-loc하는 interacting pairs 다름 (4B)
위해 probailistic establishment model 제안
Ss & CSEs 관련 inter-chromosomal interactions의 formation 설명
Depletion of MAZ reorganizes speckle-associeated inter-chromosmal interactions
MAZ depletion의 global effect를 systematically 조사
in situ Hi-C exp on MAZ-depleted HeLa cells -> HiCAN (S7F)
results
~40% genomic regions -> S-basis alteration (6D, S7G) (?)
high S-basis regions: lose S-assoc. inter-chm interaction vs low S-basis regions: gain interactions (6D, S7H)
changes in S-assoc. inter-chm interaction (6E) (?)
reproducible in KO HiC data (6E, S7K)
-> MAZ: important factor for shaping S associated inter-chm interactions both in the human & mouse
no dramatic changes in intra-chm contacts (directionality index (DI) scores (?))/ N-basis (S7I, J)
not all genomic regions were affected by MAZ depletion
hyp: effect of MAZ on S-assoc. inter-chm interactions -directly related to CSEs (S7D, E)
H3K27ac ChIP-seq on MAZ-depleted HeLa cells -> 25,022 peaks (S8A)
80 CSEs among the total 218 SEs in the control cell
only ~0.1% peaks were altered
-> the effect of MAZ depletion on inter-chm interaction: indep. to SE activity (S8B, C)
S-basis values of CSEs: more significantly decreased (vs nCSEs) (6F)
MAZ depletion specifically disrupts S-assoc. inter-chm interactions mediated by CSEs
test fnct of top 6 candidates (CSE-specific binding ptns on CSE-S assoc.)
96h KD w/ RNA interference in HeLa cells
KD effeciency <- qPCR & WB confirmed (S6 & S7A)
S assoc. of 4 CSE & 4 nCSE loci w/ 8-round DNA Oligopaint FISH
result
depletion of MAZ -> 6 target TFs와 CSE, nCSE와의 S assoc. 차이 없앰 (S7B)
높은 S-basis value 갖는 genomic regions: S assoc. 감소
낮은 S-basis value 갖는 regions: S assoc 증가
MAZ ChIP-seq -> MAZ-CSEs (vs other types of SEs) (S7D, E)
-> hyp: MAZ - crucial to establish S-CSE assoc.
functional impact of the disorganization of inter-chm interactions
RNA-seq on MAZ-depleted HeLa cells (S8D) -> 1,018 upregulated & 755 downregulated DEGs (S8E, S8, 9)
upregulated - located in increased S-basis genomic regions
downregulated: vice versa (6G)
gene ontology (GO) analysis -> downregulated: enriched in the 'cytosolic ribosome' term (6H)
S-assoc. inter-chm interactions might play an important role in the steady expr of ribosomal ptns
can't not say DEGs <-directly affacted by the MAZ depletion rather than structural alteration
S-assoc. inter-chm interactions: tightly related to the expr of a certain functional group of genes
어떤 factor가 S-CSE assoc.를 결정하는지 (underlying mechanism)
hyp: CSE (vs SSE) 에 binding하는 distinct ptn: S assoc에 관련 있을 것
CSE-specific candidate binding ptns를 indentify하기 위해 CSEs (vs SSEs & typical enhancers)에서 motif analysis함
25 ptn-binding motifs: CSEs에서 highly enriched, SSEs에서 depleted (6A)
multiple zinc finger ptns (transcription regulation & splicing) - (corr. btn each motif enrichment @ CSEs & corresponding values in S-basis (S6A)) -> 6 candidates
이 DNA-binding ptns: structurally disordered regions 포함함
aa composition: biased toward charged (R, K, D, E) & pi residues (F, Y) (S6B-I)
known to drive liquid-liquid phase separation in RNA-binding ptns
discussion