Available via license: CC BY 4.0
Content may be subject to copyright.
Cis-regulatory control of transcriptional timing and noise in response to
estrogen
Matthew Ginley-Hidinger1,2, Hosiana Abewe1,3, Kyle Osborne1,3, Katelyn L. Mortenson1,3,
Alexandra Richey1,2, Erin M. Wissink4, John Lis4, Xiaoyang Zhang1,3, Jason Gertz1,2,3,*
1. Huntsman Cancer Institute, University of Utah, Salt Lake City, UT 84112, USA
2. Department of Biomedical Engineering, University of Utah, Salt Lake City, UT 84112, USA
3. Department of Oncological Sciences, University of Utah, Salt Lake City, UT 84112, USA
4. Department of Molecular Biology and Genetics, Cornell University, Ithaca, NY 14853, USA.
* Lead contact: jay.gertz@hci.utah.edu
Abstract
Cis-Regulatory Elements (CREs) control transcription levels, temporal dynamics, and
cell-cell variation - often referred to as transcriptional noise. However, the combination of
regulatory proteins and epigenetic features necessary to control different transcription attributes
is not fully understood. Here, single-cell RNA-seq (scRNA-seq) is conducted during a time
course of estrogen treatment to identify genomic predictors of expression timing and noise. We
find that genes associated with multiple active enhancers exhibit faster temporal responses.
Synthetic modulation of enhancer activity verifies that activating enhancers accelerates
expression responses, while inhibiting enhancers results in a more gradual response. Noise is
controlled by a balance of promoter and enhancer activity. Active promoters are found at genes
with low noise levels, whereas active enhancers are associated with high noise. Finally, we
observe that co-expression across single cells is an emergent property associated with chromatin
looping, timing, and noise levels. Overall, our results indicate a fundamental tradeoff between a
gene’s ability to quickly respond to incoming signals and maintain low variation across cells.
Introduction
Cis-Regulatory elements (CREs) control the precise spatiotemporal expression of genes
across the genome. In addition to a gene’s promoter, many enhancers collaborate to control a
single gene’s expression in mammalian cells (ENCODE, 2012; Kundaje et al., 2015; Zhang et
al., 2020). External chemical signals often induce changes in cell phenotypes by altering
transcription, requiring coordinated gene expression programs. Signal transduction can lead to
transcription factor (TF) binding changes and epigenetic modifications at CREs (MacKenzie et
al., 2013). For cells to appropriately respond to stimuli, CREs must guide the amount of
transcript produced (MacKenzie et al., 2013), the timing of transcriptional changes (Kolch et al.,
2015; Wei et al., 2016), and the amount of transcriptional variation or noise (Kolch et al., 2015;
Raj and Van Oudenaarden, 2008; Raser and O'Shea, 2005). While there has been extensive
research on the role CREs play in transcription levels, less is understood about the properties of
CREs that control gene expression timing and noise.
Temporal regulation of gene expression is an essential attribute of transcriptional control
for cellular processes such as cell fate transitions (Basma et al., 2009; Chamberlain et al., 2008;
Konstantinides et al., 2022) and responses to signals (Behar and Hoffmann, 2010; Krakauer et
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
al., 2002; Uribe et al., 2021). Specific genes, often termed immediate-early genes, are rapidly
activated in response to a signal, while other genes change expression more gradually (Sheng
and Greenberg, 1990; Uhlitz et al., 2017; Uribe et al., 2021). Genes that show coordinated
trajectories are often functionally related, driving diverse phenotypes at different timescales
(Gandhi et al., 2011; Krakauer et al., 2002; Schnoes et al., 2008; Szustakowski et al., 2007).
Previous studies have identified several mechanisms that regulate transcriptional timing. One
influential factor is the state of a gene’s promoter. For example, pre-loading of RNA polymerase
II (RNAPII) at the promoter is indicative of earlier gene expression responses (Tullai et al.,
2007). Additional promoter features associated with early responding genes include TATA
motifs at the promoter, a greater number of TF binding motifs, and increased chromatin
accessibility (Murai et al., 2020; Tullai et al., 2007). Enhancers are also crucial for gene
expression timing. Inhibition or deletion of specific enhancers can prolong the time needed for a
gene to reach maximal expression without altering final expression levels (Juan and Ruddle,
2003; Simeonov et al., 2017). Stretches of potent enhancers, called super-enhancers, regulate
some immediate-early genes (Hah et al., 2015). In contrast, enhancers marked by repressive
chromatin marks, termed latent enhancers, exhibit slower activation and are associated with late-
responding genes (Ostuni et al., 2013). Generally, relatively little is known about which genomic
features in a gene’s cis-regulatory repertoire are important for influencing stimulus-dependent
temporal gene responses.
In addition to regulating gene expression timing and levels, CREs control the amount of
transcriptional noise. Transcriptional noise is a combination of intrinsic stochasticity and
extrinsic variability that cause transcript variation across a population of isogenic cells (Elowitz
et al., 2002; Fraser et al., 2021b; Kundaje et al., 2015). Cells must regulate transcriptional
variation, as both high and low variation have functional consequences. High variation can have
benefits, as cells may be more adaptable to changing environments (Pedraza et al., 2018;
Wollman, 2018) and more likely to undergo cell fate transitions (Desai et al., 2021; Suderman et
al., 2017). Noise may additionally confer the ability of a cell population to produce a diverse
output to a single incoming signal (Azpeitia et al., 2020). However, noise can be associated with
negative consequences, such as worse cancer outcomes (Han et al., 2016), cancer therapy
resistance (Qin et al., 2020; Shaffer et al., 2017), and the ability of cancer cells to metastasize
(Fidler, 1978; Nguyen et al., 2016). Both promoters and enhancers can regulate intrinsic noise
kinetics and sensitivity to extrinsic noise sources (Larsson et al., 2019). For example,
nucleosome positioning and histone modifications at the promoter are important noise regulators
(Choi and Kim, 2009; Dadiani et al., 2013; Fraser et al., 2021b; Nicolas et al., 2018; Wu et al.,
2017), with active histone marks at promoters often associated with low noise (Urban and
Johnston, 2018). Additionally, a greater number of transcription factors binding at the promoter
may be a basis for greater amounts of noise (Parab et al., 2022). The role of enhancers in
controlling mammalian expression noise is less clear. Thermodynamic modeling approaches
suggest that multiple enhancers should buffer noise (Hnisz et al., 2017), while experimental
evidence shows that super-enhancers are generally associated with noisier expression (Fraser et
al., 2021b; Wibisana et al., 2022). A remaining challenge is understanding the effect of multiple
enhancers in combination with a promoter on expression noise.
To investigate the regulatory control of timing and noise in depth, we focused on the
transcriptional response to estrogens. Estrogen Receptor ⍺ (ER) is a nuclear hormone receptor
activated by estrogens, including endogenously produced 17β-estradiol (E2). In the presence of
E2, ER becomes an active TF and regulates the expression of hundreds of genes (Bjornstrom and
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Sjoberg, 2005). ER is a clinically relevant TF, a potent oncogenic driver for endometrial and
breast cancer (Rodriguez et al., 2019a; Stanford et al., 1986), and a well-studied model TF. Upon
activation, ER both upregulates and downregulates genes at different timescales (Frasor et al.,
2003; Liberzon et al., 2015). Following an estrogen induction, ER activates successive sets of
functionally unique genes, as seen in genes related to vascularization, signaling, proliferation,
and cell cycle (Jagannathan and Robinson-Rechavi, 2011; Schnoes et al., 2008). ER has also
been shown to regulate transcriptional noise. Live cell imaging of ER target genes GREB1
(Fritzsch et al., 2018) and TFF1 (Rodriguez et al., 2019b) show that ER impacts transcriptional
noise by modulating transcriptional kinetics. The temporal, heterogeneous complexity of the ER
transcriptional program makes it an ideal model system for studying how CREs regulate
transcriptional timing and noise in response to an external stimulus.
To better understand the genomic underpinnings of transcriptional levels, timing, and
noise, we analyzed the transcriptional response to E2 using a time course of single-cell RNA-seq
(scRNA-seq) in two cell types (human breast and endometrial cancer cells). Feature ranking
approaches, using genomic data, revealed important determinants that control these
transcriptional attributes. A strong enhancer repertoire was associated with earlier changes in
gene expression, which was confirmed using functional perturbation by dCas9-based synthetic
transcription factors. Promoter features also regulate timing, such as transcriptional repressor
SIN3A being found at the promoters of “Late” genes. We uncovered a balance between
enhancers and promoters in regulating expression noise, where strong enhancers drive higher
noise and strong promoters are associated with low expression variance. The role of enhancers in
timing and noise reveals a tradeoff between expression noise and the ability to respond quickly
to incoming signals.
Results
Machine learning approach accurately predicts genomic determinants of expression levels
To uncover features of gene regulation that control levels, timing, and noise, pooled
scRNA-seq was conducted following 0-, 2-, 4-, and 8-hour E2 treatments in two cell lines:
Ishikawa (human endometrial adenocarcinoma) and T-47D (human breast carcinoma) (QC
metrics shown in Figure S1A and B). We first set out to identify determinants of mean
expression levels by focusing on the 0-hour timepoint (no E2 treatment). Integrated data from
publicly available sources (ENCODE, 2012; Shu et al., 2016; Zhang et al., 2016) and
experiments conducted for this study were quantified at promoter and enhancer regions. Due to
variations in enhancer number and strength across genes, an aggregate enhancer score was used
to capture the combined action of multiple enhancers (see methods) (Figure S1C). Genomic
features were ranked by importance for classifying low (bottom 20% of genes), medium, and
high (top 20% of genes) expression levels using the Boruta algorithm for feature selection (Kursa
and Rudnicki, 2010b), which has been previously used to uncover determinants of expression
noise in drosophila (Sigalova et al., 2020). For feature ranking, we grouped genes from both cell
types to find mutual predictors, with the expectation that there are common underlying
mechanisms for transcriptional control.
Elements of the pre-initiation complex and H3K27ac at the promoter ranked as the most
important predictors for transcript levels (Figure 1A left). These features had stronger promoter
signals for higher expressed genes (Figure 1B-E), in agreement with previous literature (Schier
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
and Taatjes, 2020; Wang et al., 2008). Important factors at promoters and enhancers showed a
general trend of increased signal for highly expressed genes (Figure 1A, right). Our dataset is
strongly biased toward activating transcription factors and activating histone marks. Plotting the
average promoter intensity compared to the average enhancer score across all confirmed datasets
verifies that strong promoters and active enhancers are associated with higher gene expression
levels (Figure 1F). Overall, the Boruta approach was successful at identifying known predictors
of transcript levels.
Figure 1. General transcription factors are the strongest predictors of gene expression levels. (A, left)
Boruta feature ranking of genomic features shows importance of a feature for predicting mean levels. (A, right)
Average signal intensity for each genomic dataset, grouped by mean expression levels is shown. Datasets shown
in bold were performed in the absence of ER activation. (B-E) Distributions of the top 4 most important ranked
features in the DMSO condition, separated by mean expression levels, show higher signal for “High” expression
groups. X-axis represents Z-scores and error bars show the mean ± 95% confidence intervals. (F) Mean enhancer
score signal for all Boruta confirmed features vs. mean promoter signal across all confirmed features is shown
where error bars show the mean ± 95% confidence intervals.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Figure 2. Feature ranking for predicting expression trajectories reveals SIN3A and multiple ERBS as
strong predictors. (A-B) UMAP dimensionality reduction plots for (A) Ishikawa and (B) T-47D show
temporal progression of cells treated with E2. Each point represents a cell and colors show timepoints post
10nM E2 induction. (C-D) Z-scores for each gene across 4 timepoints are shown for classified trajectories of
gene expression in (C) Ishikawa and (D) T-47D cells. (E) (left) Based on Boruta ranking, the top 25 most
important features are shown for classifying gene trajectories. (right) Heatmap displays the average signal by
trajectory for each predictor. Datasets shown in bold were performed in the absence of ER activation. (F-J)
Distribution of signal (Z-score) of the most important features for predicting temporal trajectories is shown.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Analysis of temporal trajectories indicates that CRE features control estrogen response
timing
scRNA-seq gene expression data following 0-, 2-, 4- and 8-hours of estrogen treatment
was used to uncover genomic predictors of temporal transcription responses to E2. Using
dimensionality-reduction UMAP plots, the temporal progression of the E2 response in single
cells can be observed (Figure 2A and B). Based on a Wilcoxon rank-sum test (Bauer, 1972),
there were approximately 2000 differentially expressed genes for each cell type between any E2
timepoint compared to the 0-hour control. scRNA-seq summed counts showed high concordance
with previously conducted bulk RNA-seq (Figure S2A and B) (Gertz et al., 2013). Compared to
bulk RNA-seq, there are more differential genes with higher expression (Figure S2A and B) and
lower fold changes (Figure S2C and D), likely due to the increased statistical power of scRNA-
seq for calling differential expression of highly expressed genes. scRNA-seq can therefore be a
valuable tool to capture subtle changes in gene expression following E2 treatment.
Based on the timepoint at which a gene is differentially regulated, genes were classified
into temporal response trajectories for up- and down-regulated genes. One class of genes rapidly
changes expression in response to E2 (termed Early genes), while another class changes more
gradually and takes longer to reach a maximum response (termed Late genes). A representative
set of Control genes was also randomly selected using stratified sampling to mirror mean levels
found in differential genes (Figure 2C and D). Early responding genes have a significant initial
response to E2 by 2 hours, then return toward baseline for both up- and down-regulation,
consistent with previous reports of pulse-like expression in immediate-early genes (Iyer et al.,
1999). In contrast, genes classified as Late show a slow and steady response over time (Figure
2C and D). As reported previously (Gertz et al., 2013), these gene expression changes were
mainly cell-type specific. However, genes with Up trajectories are more conserved between cell
types than genes with Down trajectories (Figure S2E) (p-value=0.013; t-test).
The Boruta algorithm was used to identify predictors of temporal trajectories. SIN3A
signal at the promoter was most predictive of gene expression trajectory and is associated with
Late Up genes (Figure 2E and F). MAX, which is known to repress genes in a complex with
SIN3A (Baudino and Cleveland, 2001), was also classified as important and enriched at
promoters of Late Up genes (Figure 2I). The number of ER binding sites (ERBS) that loop to the
promoter and the ER signal at enhancers were the next most important features (Figure 2E).
These two features are most significant for Early Up genes (Figure 2G and H), consistent with
previous results showing that ER is an enhancer-binding protein (Droog et al., 2016). A higher
number of enhancers is enriched for genes that respond Early (Figure S2G). However, specific
proteins (e.g., FOXA1) (Figure 2J) are more balanced between Early Up and Early Down genes
than other factors (e.g., ER) that show preferential binding to Early Up genes. Together, these
results suggest that the number of enhancers plays a prominent role in the temporal response of
genes, but transcription factors at these sites, such as ER, help control the direction of gene
expression changes. An example of an optimal decision tree was computed to examine a
potential hierarchy of factors determining a gene’s temporal response (Figure S2F). This
decision tree shows how SIN3A is the primary separator of genes into Late Up gene trajectories
and may take precedence over the ER signal. Overall, Boruta analysis of temporal trajectories
uncovers unique factors that may regulate transcription response timing and shows the
association of multiple ER-bound enhancers with a rapid up-regulation in response to E2.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Figure 3. Functional manipulation of enhancer activity alters TACSTD2 trajectory in response to
estrogen. (A) ChIP-seq, ATAC-seq, and HiChIP genome browser tracks in Ishikawa cells showing
targeted regulatory regions surrounding TACSTD2. (B-C) Expression trajectory of TACSTD2 following
estrogen induction in Ishikawa cells. (B) SID(4x)-dCas9-KRAB inhibition or (C) dCas9-VP16(10x)
activation targeted to regulatory regions temporally alters the TACSTD2 response. Controls are shown in
gray and lines represent loess regressions. (D-E) Differential of loess regressions from B and C show
expression rates of change with regulatory regions targeted by SID(4x)-dCas9-KRAB (D) or dCas9-
VP16(10x) (E). (F-G) Aggregate rates of change for all regulatory regions targeted (grey) by SID(4x)-
dCas9-KRAB (F) or dCas9-VP16(10x) (G) compared to control (black).
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Functional perturbation of CREs alters temporal responses
To test the functional relationship between CREs and transcriptional response timing,
dCas9-based activators and repressors were used to modulate the genomic activity of regulatory
regions. Gene expression responses were then measured during 8-hours of E2 treatment in
Ishikawa cells. A SID(4x)-dCas9-KRAB construct was used for repression (Carleton et al.,
2017). This construct can directly recruit SIN3A, a good predictor of Late gene expression
responses. It also recruits Histone deacetylases (HDACs) (Urrutia, 2003), corresponding to the
low H3K27ac seen at Late Up genes. For activation, dCas9-VP16(10x) was used. We have
previously shown that dCas9-VP16(10x) modulates expression from enhancers and induces
acetylation at targeted regions (Ginley-Hidinger et al., 2019). dCas9-VP16(10x) recruits many
activating cofactors, including members of the pre-initiation complex and p300, which are
associated with Early genes in the previous section.
TACSTD2 is an E2-regulated gene that is a prognostic indicator for endometrial cancer
disease-free survival (Bignotti et al., 2012), is overexpressed in some breast cancers (Shvartsur
and Bonavida, 2014), and exhibits an Early Up trajectory. Targeting the enhancers of TACSTD2,
marked by H3K27ac and ER binding (Figure 3A) with SID(4X)-dCas9-KRAB resulted in a
slower, more gradual response to E2. The time for expression to reach maximal observed levels
was increased when targeting 2 out of 3 individual enhancers (Figure 3B). When targeting
Enhancer +4.7kb, enhancer -15.2kb, and all enhancers, lower activation rates from 0 to 4 hours
were observed compared to non-targeting controls (Figure 3D). These rates were calculated
using the differential of a loess regression. Enhancer +4.7kb and enhancer -15.2kb showed
increased slope at later timepoints and similar activation levels at 8 hours compared to controls,
indicating that these enhancers can regulate the timing of a response without affecting overall
levels. Inhibition of the promoter also led to lower slopes across all timepoints and decreased
expression levels (Figure 3B and D, dark red). Synthetic activation of the same TACSTD2-linked
CREs led to a more rapid response when targeting each individual enhancer and the combination
of all enhancers (Figure 3C), as evidenced by increased activation rates between 0 and 2 hours
(Figure 3E). Again, we see that enhancer -15.2kb changes gene timing without affecting overall
transcript abundance (Figure 3E, brown), and the most substantial effect on response timing is
seen when targeting all enhancers in combination. These results imply that decreasing enhancer
activity reduces initial activation rates while activating enhancers potentiates a gene for quicker
responses to E2 (Figure 3F and G).
When targeting five putative enhancers as well as the promoter of TGFA, an Early Up
gene, with SID(4X)-dCas9-KRAB, we again observed a more gradual expression response to E2
from 3 of 5 targeted enhancers, the promoter, and a combination of all enhancers (Figure S3A).
The most substantial effects on timing were seen when targeting a distal enhancer, enhancer -
62kb, or all enhancers simultaneously (Figure S3C gray and orange). On aggregate, inhibition of
TGFA regulatory regions showed slower activation rates between early timepoints, followed by
increased rates from 6 to 8 hours relative to the control trajectory (Figure S3E). These results are
consistent with our TACSTD2 findings and indicate that decreasing enhancer activity slows the
transcriptional response.
To speed up a Late gene, dCas9-VP16(10x) was used to activate enhancers and the
promoter of PEG10. 3 out of 4 enhancers, located at +45kb, +158kb, and -305kb from the
transcription start site, led to a sharper increase in transcription at early timepoints than a control
trajectory (Figure S3B and D). Promoter targeting did not significantly speed up the response to
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
E2 but led to substantially higher expression levels (Figure S3B, dark red). On aggregate, we see
that activation of PEG10 enhancers led to earlier responses to E2, which later converge with
control rates (Figure S3F). Overall, at these three genes, the activity of promoters and individual
enhancers can control the E2 response trajectory and targeting enhancer combinations
consistently modifies temporal patterns.
Figure 4. Determinants of noise levels show a balance between active promoters driving low noise levels
and active enhancers driving high noise levels. (A, left) Boruta feature rankings shows features predictive of
noise levels. (A, right) Average signal intensity is shown by noise group for top ranked features. Datasets shown
in bold were performed in the absence of ER activation. (B-D) Distribution of signal for top ranked noise-
predicting DMSO-treated features is shown with Z-scores on the x-axis. (E) Mean enhancer score signal for all
Boruta confirmed features vs. mean promoter signal across all confirmed features for each noise level exhibits an
inverse relationship. Error bars show 95% confidence intervals. (F-G) Distribution of enhancer counts per gene,
separated by noise level, are shown for (F) Ishikawa and (G) T-47D cells.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
An enhancer-promoter dichotomy controls gene expression noise
Genes were separated into three levels of variation to find determinants of expression
noise. In scRNA-seq, low gene expression levels often have high noise due to the dropout effects
of capturing RNAs from the limited amount of RNA in a single cell. Technical variation in
scRNA-seq is related to mean levels (Brennecke et al., 2013). To examine mean-independent
noise, we used an adjusted coefficient of variation (CV), which is calculated as the residuals of a
generalized additive model (GAM) fitted to CV vs. the mean (Figure S4A). To remove any
leftover mean effects, genes were labeled as high or low noise based on whether they were in the
top 20% or bottom 20% of adjusted CV for ten different mean bins from the 0-hour timepoint of
both cell types.
The strongest predictors of noise levels were SIN3A and JUN at the promoter, both
associated with low noise (Figure 4A-B,D). Generally, a strong promoter signal was related to
low noise across features, with some exceptions, such as p300 (Figure 4A, right panel). Most
enhancer features were associated with high noise, with ER and FOXA1 at enhancers being the
most predictive (Figure 4A). Another feature scored as highly important was tri-methylation at
histone H3 lysine 4 (H3K4me3), which is strongly associated with low amounts of noise (Figure
4C), supporting a role for H3K4me3 in controlling noise. This result is consistent with a previous
study that found a relationship between H3K4me3 breadth and transcriptional consistency
(Benayoun et al., 2014).
These results motivated the broader evaluation of how noise relates to promoter and
enhancer activity. Analysis of the average promoter intensity across all confirmed datasets and
the average enhancer score revealed an inverse relationship between promoters and enhancers
(Figure 4E). Genes with high noise levels had high enhancer scores and low promoter signals.
Conversely, genes with low noise levels had low enhancer scores and high promoter signals. To
confirm this relationship in a third cell type, we analyzed publicly available scRNA-seq and
genomic data from LNCaP cells, a prostate cancer cell line. The same association was observed
between enhancer-driven gene regulation and higher noise (Figure S4B). Consistent with these
findings, more enhancers connected to a gene is associated with greater noise (Figure 4F and G).
Several features are associated with noise and either expression levels or timing
Levels, noise, and timing analysis showed different importance rankings for genomic
features. Hierarchical clustering of these features by importance score revealed five major
clusters (Figure 5A). The largest clusters consisted of features specific to each analysis. Two
smaller clusters were composed of factors necessary for both mean and noise or both trajectory
and noise. Notably missing were factors important for both mean levels and temporal
predictions. In general, the importance score from the Boruta algorithm is more similar between
noise and mean or noise and trajectory compared to mean and trajectory, as seen from the first
two PCA dimensions calculated from the feature importance matrix (Figure 5B) and the
correlation between importance scores (Figure S5B-D). The relationship between noise and our
other analyses suggests that noise may be an intermediary between baseline levels and temporal
regulation and that mean levels do not strongly influence response trajectory.
By comparing the ratio of promoter to enhancer features in each cluster, we see that mean
levels are the most promoter driven. Noise and trajectory utilize promoter and enhancer features
more evenly (Figure 5C). Enhancer features important for predicting mean levels are also likely
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
to predict noise. Together, these results indicate that enhancers are more critical for noise and
trajectory and that many genomic signals preferentially predict either levels, noise, or trajectory.
Boruta importance scores do not capture directionality. To examine which features are associated
with classification groups for each analysis, the group with the maximum signal is shown in
Figure S5A. Promoter features are generally associated with high mean levels and low noise.
Enhancer features are also associated with increased mean levels, but contrary to promoters, they
show a higher signal for high noise levels. Enhancer scores are almost always the highest for
Early Up trajectories (Figure S5A). Our results suggest that active enhancers drive high noise
and rapid up-regulation in response to E2, while promoters consistently drive low noise and high
mean expression.
Co-expression of genes is based on looping, timing, and noise levels
scRNA-seq offers a unique advantage in studying the co-regulation of genes and the
possible mechanisms that underlie co-regulation. Using the H3K27ac HiChIP data, we found that
looping can affect co-expression in several ways. First, we found that genes whose promoters
loop together correlate significantly more than groups of randomly paired control genes at the 0-
hour timepoint (Figure 6A and B). Genes whose promoters both loop to a shared enhancer are
significantly more correlated across single cells (Figure 6C and D). These results indicate that
the 3D genome structure may be involved or associated with gene co-expression across single
cells.
We next evaluated co-expression during the E2 treatment time course. Co-expression was
measured using pairwise spearman correlation in single cells. In Ishikawa cells, both Early Up
and Early Down genes show increasing pairwise co-expression over time (Figure 6E). Genes that
respond late exhibited less change in correlation, with Late Up genes increasing correlation
slightly by 8 hours and Late Down genes slightly decreasing in correlation. In T-47D cells, we
see the most significant increase in co-correlation at 2 hours for Early Up and Early Down genes
(Figure 6F). Late Up and Late Down genes show slight increases in correlation during the time
course.
The levels of noise also change the probability of two genes being correlated. Perhaps
expected, genes with high noise levels also show a broader distribution in their pairwise
correlations, resulting in genes with high noise being more likely to have extremely correlated or
anti-correlated expression with each other than low noise genes (Figure 6G and H). These results
indicate that 3D interactions, control of temporal trajectory, and noise regulation can impact gene
co-regulation.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Figure 5. Importance comparison shows that mean and trajectory are regulated by distinct genomic
features. (A) Heatmap shows importance scores from each analysis type, normalized by column, and scaled by
row. Datasets shown in bold were performed in the absence of ER activation. (B) PCA plot based on importance
scores show the relationship of importance scores for mean levels, noise, and trajectory. Percentages denote
percent of variance explained by each principal component. (C) Proportion of features associated with enhancers,
promoters, or features not associated with a specific regulatory element type, are plotted by cluster.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Figure 6. Co-expression
changes are observed
based on looping,
trajectory, and levels of
noise. (A-B) Pairs of genes
with promoters that loop to
one another are
significantly more
correlated across cells at the
0-hour timepoint than
randomly selected gene
pairs for Ishikawa (A) and
T-47D (B). Bonferroni
adjusted Wilcoxon p-values
are shown with respect to
control. (C-D) Pairs of
genes with a shared
enhancer are more
correlated than randomly
paired genes for Ishikawa
(C) and T-47D (D).
Wilcoxon p-values are
shown with respect to
control. (E-F) Pairwise
spearman correlation for
genes within different
trajectories is shown.
Significance values show
Wilcoxon p-values of the 8-
hour (E) or the 2-hour (F)
timepoint with respect to
the 0-hour timepoint. (G-H)
Range of pairwise
correlations for high noise
levels is greater than the
range for pairs of low noise
genes in Ishikawa (G) and
T-47D (H). (left panel)
Distribution of Spearman
pairwise correlations for
genes with high and low
noise. (right panel)
Spearman correlations were
grouped into quantiles and
bars show proportion at
each quantile that are pairs
of low or high noise genes.
Significance values for all
subpanels are as follows: (*
p < 0.05; ** p < 1x10-5;
*** p < 1x10-10; **** p <
1x10-15).
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Discussion
To investigate the genomic underpinnings of the temporal transcriptional response to
estrogen, we conducted scRNA-seq at several timepoints. scRNA-seq was able to capture more
subtle changes in gene expression of highly expressed genes compared to bulk RNA-seq due to
the increased statistical power. Using a feature ranking approach, we identified several features
associated with E2 response timing, including SIN3A signal at the promoter of “Late Up” genes
and more ER-bound enhancers regulating “Early Up” genes. In general, multiple enhancers are
more predictive of “Early” gene trajectories. Functional evaluation of enhancers revealed that
multiple enhancers regulate timing at each gene tested and that enhancers, in combination, are
consistent regulators of expression trajectory. We conclude that an active enhancer repertoire is
necessary for rapid gene responses. Active enhancers may present chromatin that is more open to
ER binding. Alternatively, other TFs already present at enhancers could stabilize the binding of
ER, permitting ER to activate gene expression immediately. Contrastingly, SIN3A and MAX at
the promoter, known to repress gene expression together (Baudino and Cleveland, 2001), slow a
gene’s response to E2, even when a gene is associated with strong ER-bound enhancers.
Activation of gene expression may first require removing repressive signals at the promoter,
explaining the more gradual responses. A similar mechanism has been described in an enhancer
context, where inactive enhancers must first be activated by transcription factors before
activation of gene expression can occur, causing more gradual gene expression changes (Ostuni
et al., 2013).
Genomic data analysis revealed that opposing enhancer and promoter activities control
expression noise. Active promoters are associated with low noise levels, whereas multiple active
enhancers are associated with high noise levels. In concordance with this observation, synthetic
activation of promoters drives lower noise levels at several genes (Fraser et al., 2021a).
Additionally, activation of multiple enhancers causes high noise at the NF-κB locus (Wibisana et
al., 2022). Our results support a unified model where a balance between enhancers and
promoters control noise. Both intrinsic and extrinsic noise could potentially explain the observed
noise distributions (Ham et al., 2021). If intrinsic noise is the driving factor, we expect promoters
to cause high-frequency, near-constant transcription levels and enhancers to cause infrequent,
high-amplitude bursts of expression (Raj and Van Oudenaarden, 2008). Noise caused by CREs
could also be due to extrinsic noise. Promoters may lead to low noise, as fluctuations in upstream
factors may be insignificant compared to activation by an ensemble of transcription factors
bound to the promoter. In contrast, enhancers may drive higher noise levels by increasing
sensitivity to upstream elements. Our above observation that enhancers drive rapid temporal
responses further supports the theory that strong enhancers increase a gene’s sensitivity to
incoming signals.
For a gene to respond quickly to a signal, it must be sensitized to incoming signals, which
may inherently drive higher levels of noise. A noise-robustness tradeoff has been proposed
previously when observing changes in gene expression over developmental time in drosophila
(Sigalova et al., 2020) and in a mathematical framework that showed variations are necessary for
a gene’s responsiveness (Boe et al., 2022). However, a regulatory mechanism has not been
found. Our results point to multiple enhancers being a primary genomic feature associated with
both high expression noise and rapid response timing. Additionally, we found that a strong
promoter is likely to cause more robust gene expression but limited responsiveness. This tradeoff
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
begs the question of whether specific genes can respond quickly to a signal without high cell-cell
variation and what mechanisms may be in place to prevent noise in this case.
Co-expression analysis of gene pairs showed that co-expression properties depend on
looping, timing, and noise. We found that genes with shared enhancers and looped promoters
correlate more in individual cells, genes with different trajectories correlate differently over time,
and genes with high noise levels are more likely to be strongly correlated or anticorrelated. These
observations have considerable implications for gene regulatory networks (GRNs) in single cells,
as co-expression often underlies regulatory networks. Dynamic adjustment of regulatory
networks may have critical functional outcomes for a cell population (Borriello et al., 2020). For
example, GRNs which confer resistance to therapeutics may occur at distinct timepoints
following treatment (Zhang et al., 2019). Our results indicate that genes with high noise may lead
to a broader range of implemented regulatory networks across single cells, enhancing cellular
heterogeneity. Further studies into functional GRNs are warranted to determine how noise and
timing affect single-cell phenotypes through the co-expression of many genes. Overall, our study
shows that enhancers and promoters can play distinct roles in the timing and variation of a
transcriptional response.
Methods
Cell culture
T-47D and Ishikawa cells were cultured in RPMI 1640 medium (Gibco) with 10% fetal
bovine serum (Gibco) and 1% penicillin–streptomycin (Gibco). LNCaP cells were cultured in
RPMI media with 10% FBS supplemented. Cells were incubated at 37°C with 5% CO2. 5 days
before estrogen inductions, cells were transferred to hormone-depleted media consisting of
phenol red-free RPMI (Gibco) with 10% charcoal-dextran stripped fetal bovine serum (Sigma-
Aldrich) and 1% penicillin–streptomycin (Gibco).
ChIP-seq
After 5 days in hormone-depleted media, cells were plated in 15cm dishes at
approximately 60% confluency 1 day before estrogen induction. Cells were treated with vehicle
(DMSO) or E2 at a final concentration of 10nM for either 1 hour for transcription factor ChIP-
seqs, or 8 hours for histone marker ChIP-seqs. ChIP and library preparation was performed as
previously described (Reddy et al., 2009). Antibodies used for this study were MAX (Sant Cruz
sc-8011), LSD1 (abcam ab 17721), TAF1 (sc-735), c-MYC (Santa Cruz sc-40), H3K4me3 (Cell
Signaling 9751S), H3K4me1 (Cell Signaling 5326S), SIN3A (produced as previously described)
(Hassig et al., 1997), RARA (Santa Cruz sc-515796) and JUN (BD Biosciences
558036). Libraries were sequenced using either an Illumina HiSeq 2500 or Illumina NovaSeq
6000 as single- or paired-end 50 bp reads, then aligned to hg19 using bowtie with parameters -m
1 –t –best -q -S -l 32 -e 80 -n 2 (Langmead et al., 2009). Signal intensity was extracted from bam
files using samtools view with parameter -c (Li et al., 2009). In the cases where peaks were
called, peak calling was done using Macs2 with the default q-value cutoff of 0.05 and mfold ratio
between 15 and 100 (Zhang et al., 2008).
H3K27ac HiChIP
HiChIP experiments were performed as previously described (Mumbach et al., 2016)
using an antibody that recognizes H3K27ac (Abcam, ab4729). Ishikawa cells were treated with
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
either 10 nM E2 for 1 hour or DMSO as a vehicle control. HiChIP in Ishikawa cells was
conducted using restriction enzyme DpnII (New England Biolabs). Crosslinked chromatin was
sonicated using an EpiShear probe-in sonicator (Active Motif) with three cycles of 30 seconds at
an amplitude of 40% with 30 seconds rest between cycles. HiChIP libraries were sequenced on
NovaSeq 6000 as paired end 50 base pair reads to an average depth of 300–400 million read-
pairs per sample.
Experiments in T-47D and LNCaP cells were conducted using restriction enzyme MboI
(New England Biolabs). Crosslinked chromatin was sonicated using Covaris E220 with the
settings of fill level=10, duty cycle=5, PIP=140, cycles per burst=200, time=4 mins. HiChIP
libraries were sequenced on HiSeq 2500 as paired end 75 base pair reads to ~50 million read
pairs per sample.
Reads were aligned to human hg19 reference genome using HiC-Pro (Servant et al.,
2015). Hichipper (Lareau and Aryee, 2018) was used to perform restriction site bias-aware
modeling of the output from HiC-Pro and to call interaction loops. In Ishikawa cells, DMSO and
E2 treated HiChIP loops were combined to identify all possible putative enhancers. In all
datasets, loops with less than 3 reads or FDR >= .05 were filtered out.
PRO-seq
PRO-seq libraries were generated as described in Mahat et al., 2016 (Mahat et al., 2016).
Briefly, Ishikawa and T-47D cells were grown in hormone-depleted RPMI for five days, then
2x106 cells were plated into two 10 cm dishes per condition with RPMI lacking phenol red
supplemented with 10% charcoal/dextran-stripped FBS and penicillin. Cells were treated with
vehicle (DMSO) or 10 nM E2 for 45 minutes, then permeabilized for five minutes with
permeabilization buffer [10 mM Tris-HCl, pH 7.4; 300 mM sucrose; 10 mM KCl; 5 mM MgCl2;
1 mM EGTA; 0.05% Tween-20; 0.1% NP40 substitute; 0.5 mM DTT, protease inhibitor cocktail
ml(Roche); and SUPERaseIn RNase Inhibitor (Ambion)]. The nuclear run-on was performed by
adding permeabilized cells to run-on mixture [final composition was 5 mM Tris, pH 8.0; 25 mM
MgCl2; 0.5 mM DTT; 150 mM KCl; 200 μM rATP; 200 μM rGTP; 20 μM biotin-11-rCTP
(Perkins Elmer); 20 μM biotin-11-rUTP (Perkins Elmer); 1 U/μL SUPERase In RNase Inhibitor
(Ambion); 0.5% Sarkosyl], then incubating at 37°C for 5 minutes. RNA was extracted with
Trizol LS (Ambion), fragmented with 0.2 N NaOH for 8 minutes on ice, then neutralized with
0.5 M Tris, pH 6.8, followed by buffer exchange with a P-30 column (Bio-Rad). Biotinylated
RNAs were enriched with Dynabeads M280 Streptavidin (Invitrogen), then RNA was extracted
with Trizol (Ambion), followed by 3′ adapter ligation using T4 RNA ligase (NEB). Biotinylated
RNAs were enriched for a second time, followed by 5′ cap repair with RppH (NEB) and 5′
hydroxyl repair with PNK (NEB). The 5′ adapter was ligated with T4 RNA ligase (NEB),
followed by a third biotinylated RNA enrichment. Reverse transcription was performed with the
RP1 primer. Samples were PCR amplified for 13 cycles, then cleaned up with Agencourt
AMPure XP beads (Beckman Coulter). Libraries were sequenced on an Illumina HiSeq 2500,
generating a 50nt read. Reads were processed using cutadapt (Martin, 2011) with parameters -a
TGGAATTCTCGGGTGCCAAGG --cut 7 --length 42 -m 21. Reverse complement sequences
were generated using fastx_reverse_complement from the FASTX toolkit (v 0.0.13) (Hannon,
2010). Reads were then aligned to hg19 with bowtie2 (Langmead and Salzberg, 2012) in end-to-
end mode, and non-uniquely aligned reads were discarded.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
scRNA-seq
Cells were treated with 10nM E2 for 0 (vehicle treated), 2, 4, and 8 hours. To mitigate
technical batch effects, cells were labeled via MULTI-seq as previously described (McGinnis et
al., 2019). Cells from different timepoints were mixed and then prepared according to the 10x
Genomics sample prep user guide (Single Cell Suspensions from Cultured Cell Lines for Single
Cell RNA Sequencing, 2017). Cells were separated into single-cell emulsions using the 10x
Genomics Chromium Controller with a targeted recovery of 10,000 cells. Sequencing libraries
were prepared using the 10X Genomics Next GEM Single Cell 3' Gene Expression Library prep
v3.1. Sequencing was performed on an Illumina NovaSeq 6000 with 150bp read length.
Sequencing output was processed from reads to counts using the 10x Genomics Cell Ranger
v3.1.0 pipeline. MULTI-seq calls were processed using the demultiplex R package (McGinnis et
al., 2019) and mapped back to the E2 timepoints. Counts were log normalized using the Seurat
v3 R package (Stuart et al., 2019), then filtered using custom cutoffs (Figure S1A and B). Genes
are filtered to have a mean greater than 0.01 across all timepoints.
scRNA-seq analysis: classification of trajectory and noise levels
Computational analysis of trajectory and noise levels were conducted using R (Team,
2013). Trajectory classification was done using a Wilcoxon test (Bauer, 1972) to find genes
whose single-cell distributions significantly change at different timepoints compared to the 0hr
timepoint. Genes that change significantly by 2 hours are classified as either “Early Up” or
“Early Down”. Genes with changes seen at 4 or 8 hours are classified as “Late”. To select
control genes with similar mean distributions to those genes that are regulated by E2, we used a
stratified sampling approach to select control genes that are not significantly regulated.
Our noise metric is defined as the residuals from a generalized additive model (GAM)
regression fitted to the CV vs mean for all genes. Regression was performed on log2(CV + 1) vs
mean curve using the gam function from the mgcv R package with formula y ~ s(x, bs = "cs")
(Wood, 2006). Residuals were then transformed back to the original scale. Noise levels were
determined using the GAM-adjusted CV. To account for the different scales of noise in different
mean levels, genes were binned by mean into 3 groups by quantile. The top 20% and bottom
20% of genes in each quantile were labeled as “High” and “Low” noise, respectively.
Feature importance analysis
Promoters were defined as 500bp regions centered on the transcriptional start site, as
annotated in the RefSeq database (Pruitt et al., 2013). Enhancers were called using H3K27ac
HiChIP data and H3K27ac ChIP-seq peaks. Enhancers were defined as H3K27ac peaks within
HiChIP anchors that loop to the promoter. Integrated signal for each promoter and enhancer was
collected from all datasets using samtools view -c (Li et al., 2009). Z-scores were calculated
across all genes for input to feature ranking algorithms. An enhancer score was calculated to
account for signal at multiple enhancers, using the formula
"
𝑙𝑜𝑔
!
(
𝑠 + 1
)
"
#
where n represents the number of enhancers associated with a gene and s represents the z-score
of integrated genomic signal at each enhancer.
Number of enhancers was defined as the number of H3K27ac peaks which loop to the
promoter, as determined by HiChIP. Number of ERBS was calculated as the number of
enhancers that overlapped with ER ChIP-seq peaks. Feature ranking was performed using the
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Boruta package in R (Kursa and Rudnicki, 2010a) with default parameters and 100 maximum
iterations. Gene length was calculated from RefSeq transcript annotations (Pruitt et al., 2013).
An example decision tree was determined using the rpart function with parameters
minbucket=50 and cp=0.007 (Breiman et al., 2017). Average enhancer score and promoter signal
was calculated using “confirmed” variables from Boruta analysis. The average of Z-score signal
for confirmed variables was taken for all variables associated with either the promoter or
enhancer, not including number of enhancers, number of ERBS, or gene length.
Generation of stable dCas9-VP16(10x) cell lines
Ishikawa cells were plated in 6-well plates at 60% confluency. Cells were transfected
with Addgene plasmid 48227 (a gift from Rudolf Jaenisch) (Cheng et al., 2013) containing
dCas9-VP16(10x) with a P2A linker and neomycin resistance gene. Fugene HD (Promega) was
used for transfection at a 3:1 reagent:DNA ratio. dCas9-VP16(10x) plasmid was linearized with
restriction enzyme AflII (New England Biolabs R0520S). Successful integration of the dCas9-
VP16(10x) plasmid was selected for using G418 (Thermo Scientific) at a concentration of 800
µg/mL for approximately 2 weeks. Successful expression of the dCas9 plasmid was verified
using qPCR for dCas9 as well as qPCR for successful activation of a control gene, IL1RN
(Figure S3G). Cells were then maintained at a lower concentration of 400 µg/ml G418.
gRNA design and transfection
gRNAs were designed using the Benchling gRNA design tool (Benchling, 2020). 4
gRNAs were designed per targeted region. gRNAs were cloned into plasmids as previously
described (Carleton et al., 2017). gRNA sequence and adjacent PAM are listed in Table S1. Prior
to transfection, Ishikawa cells were plated in 48-well plates at 80,000 cells/well. 24 hours after
plating, gRNAs were transfected into cell using Fugene HD (Promega) at a manufacturer
suggested 3:1 reagent:DNA ratio. gRNA transfection was selected for using 1 µg/mL puromycin.
8-hour E2 time courses were started roughly 24 hours after addition of puromycin.
RNA isolation and qPCR gene expression analysis
Cells were lysed with Buffer RLT Plus (Qiagen) containing 1% beta-mercaptoethanol
(Sigma). RNA was purified using the ZR-96-well Quick-RNA kit (Zymo Research). Gene
expression was measured using qPCR with reagents from the Power SYBR Green RNA-to-Ct 1-
step kit (Applied Biosystems), 50ng RNA per reaction, and 40 cycles on a CFX Connect light
cycler (BioRad). qPCR primers are listed in Table S2. Relative expression was calculated using
the ΔΔCt method with CTCF as a reference gene. Best fit lines were determined using the loess
function in R (Cleveland et al., 2017) and formula y ~ x. Differential rates of change were
calculated by determining the average slope of the loess fit for each 1-hour window.
Author Contributions
Conceptualization, M.G-H. and J.G.; Methodology, M.G-H. and J.G.; Investigation,
M.G-H., H.A., K.L.M., and E.M.W; Formal Analysis, M.G-H., J.G., and X.Z.; Writing - Original
Draft, M.G-H. and J.G.; Writing – Review & Editing, All authors; Supervision, J.G., J.L., and
X.Z.; Funding Acquisition, J.G.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Acknowledgements
Funding for this work came from National Institute of Health (NIH)/National Human
Genome Research Institute (NHGRI) R01 HG008974 to J.G. and the Huntsman Cancer Institute.
Research reported in this publication utilized the High-Throughput Genomics Shared Resource
at the University of Utah and was supported by NIH/National Cancer Institute (NCI) award P30
CA042014. We thank Craig M. Rush for experimental guidance, Jeffery Vahrenkamp for
analysis advice, and Gertz lab members for their suggestions on the study and manuscript.
Data Availability
ChIP-seq, HiChIP, PRO-seq, and scRNA-seq data are available at the Gene Expression
Omnibus under accession GSE227245.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
References
Azpeitia, E., Balanzario, E.P., and Wagner, A. (2020). Signaling pathways have an inherent need
for noise to acquire information. BMC Bioinformatics 21, 462. 10.1186/s12859-020-03778-x.
Basma, H., Soto-Gutiérrez, A., Yannam, G.R., Liu, L., Ito, R., Yamamoto, T., Ellis, E., Carson,
S.D., Sato, S., Chen, Y., et al. (2009). Differentiation and transplantation of human embryonic
stem cell-derived hepatocytes. Gastroenterology 136, 990-999. 10.1053/j.gastro.2008.10.047.
Baudino, T.A., and Cleveland, J.L. (2001). The Max network gone mad. Mol Cell Biol 21, 691-
702. 10.1128/mcb.21.3.691-702.2001.
Bauer, D.F. (1972). Constructing confidence sets using rank statistics. Journal of the American
Statistical Association 67, 687-690.
Behar, M., and Hoffmann, A. (2010). Understanding the temporal codes of intra-cellular signals.
Current Opinion in Genetics & Development 20, 684-693.
https://doi.org/10.1016/j.gde.2010.09.007.
Benayoun, B.A., Pollina, E.A., Ucar, D., Mahmoudi, S., Karra, K., Wong, E.D., Devarajan, K.,
Daugherty, A.C., Kundaje, A.B., Mancini, E., et al. (2014). H3K4me3 breadth is linked to cell
identity and transcriptional consistency. Cell 158, 673-688. 10.1016/j.cell.2014.06.027.
Benchling (2020). Benchling.
Bignotti, E., Zanotti, L., Calza, S., Falchetti, M., Lonardi, S., Ravaggi, A., Romani, C.,
Todeschini, P., Bandiera, E., Tassi, R.A., et al. (2012). Trop-2 protein overexpression is an
independent marker for predicting disease recurrence in endometrioid endometrial carcinoma.
BMC Clinical Pathology 12, 22. 10.1186/1472-6890-12-22.
Bjornstrom, L., and Sjoberg, M. (2005). Mechanisms of estrogen receptor signaling:
convergence of genomic and nongenomic actions on target genes. Molecular endocrinology 19,
833-842.
Boe, R.H., Ayyappan, V., Schuh, L., and Raj, A. (2022). Allelic correlation is a marker of trade-
offs between barriers to transmission of expression variability and signal responsiveness in
genetic networks. Cell Systems 13, 1016-1032. e1016.
Borriello, E., Walker, S.I., and Laubichler, M.D. (2020). Cell phenotypes as macrostates of the
GRN dynamics. J Exp Zool B Mol Dev Evol 334, 213-224. 10.1002/jez.b.22938.
Breiman, L., Friedman, J.H., Olshen, R.A., and Stone, C.J. (2017). Classification and regression
trees (Routledge).
Brennecke, P., Anders, S., Kim, J.K., Kołodziejczyk, A.A., Zhang, X., Proserpio, V., Baying, B.,
Benes, V., Teichmann, S.A., Marioni, J.C., and Heisler, M.G. (2013). Accounting for technical
noise in single-cell RNA-seq experiments. Nature Methods 10, 1093-1095. 10.1038/nmeth.2645.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Carleton, J.B., Berrett, K.C., and Gertz, J. (2017). Multiplex Enhancer Interference Reveals
Collaborative Control of Gene Regulation by Estrogen Receptor α-Bound Enhancers. Cell Syst
5, 333-344.e335. 10.1016/j.cels.2017.08.011.
Chamberlain, C.E., Jeong, J., Guo, C., Allen, B.L., and McMahon, A.P. (2008). Notochord-
derived Shh concentrates in close association with the apically positioned basal body in neural
target cells and forms a dynamic gradient during neural patterning. Development 135, 1097-
1106. 10.1242/dev.013086.
Cheng, A.W., Wang, H., Yang, H., Shi, L., Katz, Y., Theunissen, T.W., Rangarajan, S.,
Shivalila, C.S., Dadon, D.B., and Jaenisch, R. (2013). Multiplexed activation of endogenous
genes by CRISPR-on, an RNA-guided transcriptional activator system. Cell research 23, 1163-
1171.
Choi, J.K., and Kim, Y.-J. (2009). Intrinsic variability of gene expression encoded in nucleosome
positioning sequences. Nature Genetics 41, 498-503. 10.1038/ng.319.
Cleveland, W.S., Grosse, E., and Shyu, W.M. (2017). Local regression models. In Statistical
models in S, (Routledge), pp. 309-376.
Dadiani, M., Van Dijk, D., Segal, B., Field, Y., Ben-Artzi, G., Raveh-Sadka, T., Levo, M.,
Kaplow, I., Weinberger, A., and Segal, E. (2013). Two DNA-encoded strategies for increasing
expression with opposing effects on promoter dynamics and transcriptional noise. Genome
Research 23, 966-976. 10.1101/gr.149096.112.
Desai, R.V., Chen, X., Martin, B., Chaturvedi, S., Hwang, D.W., Li, W., Yu, C., Ding, S.,
Thomson, M., Singer, R.H., et al. (2021). A DNA repair pathway can regulate transcriptional
noise to promote cell fate transitions. Science 373, eabc6506. doi:10.1126/science.abc6506.
Droog, M., Mensink, M., and Zwart, W. (2016). The Estrogen Receptor α-Cistrome Beyond
Breast Cancer. Mol Endocrinol 30, 1046-1058. 10.1210/me.2016-1062.
Elowitz, M.B., Levine, A.J., Siggia, E.D., and Swain, P.S. (2002). Stochastic Gene Expression in
a Single Cell. Science 297, 1183-1186. doi:10.1126/science.1070919.
ENCODE, C. (2012). An integrated encyclopedia of DNA elements in the human genome.
Nature 489, 57-74. 10.1038/nature11247.
Fidler, I.J. (1978). Tumor heterogeneity and the biology of cancer invasion and metastasis.
Cancer research 38, 2651-2660.
Fraser, L.C., Dikdan, R.J., Dey, S., Singh, A., and Tyagi, S. (2021a). Reduction in gene
expression noise by targeted increase in accessibility at gene loci. Proceedings of the National
Academy of Sciences 118, e2018640118.
Fraser, L.C.R., Dikdan, R.J., Dey, S., Singh, A., and Tyagi, S. (2021b). Reduction in gene
expression noise by targeted increase in accessibility at gene loci. Proceedings of the National
Academy of Sciences 118, e2018640118. doi:10.1073/pnas.2018640118.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Frasor, J., Danes, J.M., Komm, B., Chang, K.C.N., Lyttle, C.R., and Katzenellenbogen, B.S.
(2003). Profiling of Estrogen Up- and Down-Regulated Gene Expression in Human Breast
Cancer Cells: Insights into Gene Networks and Pathways Underlying Estrogenic Control of
Proliferation and Cell Phenotype. Endocrinology 144, 4562-4574. 10.1210/en.2003-0567.
Fritzsch, C., Baumgärtner, S., Kuban, M., Steinshorn, D., Reid, G., and Legewie, S. (2018).
Estrogen‐dependent control and cell‐to‐cell variability of transcriptional bursting. Molecular
Systems Biology 14, e7678. 10.15252/msb.20177678.
Gandhi, S.J., Zenklusen, D., Lionnet, T., and Singer, R.H. (2011). Transcription of functionally
related constitutive genes is not coordinated. Nat Struct Mol Biol 18, 27-34. 10.1038/nsmb.1934.
Gertz, J., Savic, D., Varley, E., Katherine, Partridge, C., E., Safi, A., Jain, P., Cooper, M.,
Gregory, Reddy, E., Timothy, Crawford, E., Gregory, and Myers, M., Richard (2013). Distinct
Properties of Cell-Type-Specific and Shared Transcription Factor Binding Sites. Molecular Cell
52, 25-36. 10.1016/j.molcel.2013.08.037.
Ginley-Hidinger, M., Carleton, J.B., Rodriguez, A.C., Berrett, K.C., and Gertz, J. (2019).
Sufficiency analysis of estrogen responsive enhancers using synthetic activators. Life Science
Alliance 2, e201900497. 10.26508/lsa.201900497.
Hah, N., Benner, C., Chong, L.-W., Yu, R.T., Downes, M., and Evans, R.M. (2015).
Inflammation-sensitive super enhancers form domains of coordinately regulated enhancer RNAs.
Proceedings of the National Academy of Sciences 112, E297-E302.
doi:10.1073/pnas.1424028112.
Ham, L., Jackson, M., and Stumpf, M.P. (2021). Pathway dynamics can delineate the sources of
transcriptional noise in gene expression. Elife 10. 10.7554/eLife.69324.
Han, R., Huang, G., Wang, Y., Xu, Y., Hu, Y., Jiang, W., Wang, T., Xiao, T., and Zheng, D.
(2016). Increased gene expression noise in human cancers is correlated with low p53 and
immune activities as well as late stage cancer. Oncotarget 7, 72011-72020.
10.18632/oncotarget.12457.
Hannon, G.J. (2010). FASTX-Toolkit FASTQ/A short-reads pre-processing tools,
http://hannonlab.cshl.edu/fastx_toolkit/.
Hassig, C.A., Fleischer, T.C., Billin, A.N., Schreiber, S.L., and Ayer, D.E. (1997). Histone
deacetylase activity is required for full transcriptional repression by mSin3A. Cell 89, 341-347.
10.1016/s0092-8674(00)80214-7.
Hnisz, D., Shrinivas, K., Young, R.A., Chakraborty, A.K., and Sharp, P.A. (2017). A Phase
Separation Model for Transcriptional Control. Cell 169, 13-23. 10.1016/j.cell.2017.02.007.
Iyer, V.R., Eisen, M.B., Ross, D.T., Schuler, G., Moore, T., Lee, J.C., Trent, J.M., Staudt, L.M.,
Hudson, J., Jr., Boguski, M.S., et al. (1999). The transcriptional program in the response of
human fibroblasts to serum. Science 283, 83-87. 10.1126/science.283.5398.83.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Jagannathan, V., and Robinson-Rechavi, M. (2011). Meta-analysis of estrogen response in MCF-
7 distinguishes early target genes involved in signaling and cell proliferation from later target
genes involved in cell cycle and DNA repair. BMC systems biology 5, 1-14.
Juan, A.H., and Ruddle, F.H. (2003). Enhancer timing of Hox gene expression: deletion of the
endogenous Hoxc8 early enhancer. Development 130, 4823-4834. 10.1242/dev.00672.
Kolch, W., Halasz, M., Granovskaya, M., and Kholodenko, B.N. (2015). The dynamic control of
signal transduction networks in cancer cells. Nature Reviews Cancer 15, 515-527.
10.1038/nrc3983.
Konstantinides, N., Holguera, I., Rossi, A.M., Escobar, A., Dudragne, L., Chen, Y.-C., Tran,
T.N., Martínez Jaimes, A.M., Özel, M.N., Simon, F., et al. (2022). A complete temporal
transcription factor series in the fly visual system. Nature 604, 316-322. 10.1038/s41586-022-
04564-w.
Krakauer, D.C., Page, K.M., and Sealfon, S. (2002). Module dynamics of the GnRH signal
transduction network. J Theor Biol 218, 457-470.
Kundaje, A., Meuleman, W., Ernst, J., Bilenky, M., Yen, A., Heravi-Moussavi, A., Kheradpour,
P., Zhang, Z., Wang, J., Ziller, M.J., et al. (2015). Integrative analysis of 111 reference human
epigenomes. Nature 518, 317-330. 10.1038/nature14248.
Kursa, M.B., and Rudnicki, W.R. (2010a). Feature selection with the Boruta package. Journal of
statistical software 36, 1-13.
Kursa, M.B., and Rudnicki, W.R. (2010b). Feature Selection with the Boruta Package. Journal of
Statistical Software 36, 1 - 13. 10.18637/jss.v036.i11.
Langmead, B., and Salzberg, S.L. (2012). Fast gapped-read alignment with Bowtie 2. Nature
methods 9, 357-359.
Langmead, B., Trapnell, C., Pop, M., and Salzberg, S.L. (2009). Ultrafast and memory-efficient
alignment of short DNA sequences to the human genome. Genome biology 10, 1-10.
Lareau, C.A., and Aryee, M.J. (2018). hichipper: a preprocessing pipeline for calling DNA loops
from HiChIP data. Nature methods 15, 155-156.
Larsson, A.J.M., Johnsson, P., Hagemann-Jensen, M., Hartmanis, L., Faridani, O.R., Reinius, B.,
Segerstolpe, Å., Rivera, C.M., Ren, B., and Sandberg, R. (2019). Genomic encoding of
transcriptional burst kinetics. Nature 565, 251-254. 10.1038/s41586-018-0836-1.
Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G.,
Durbin, R., and Subgroup, G.P.D.P. (2009). The Sequence Alignment/Map format and
SAMtools. Bioinformatics 25, 2078-2079. 10.1093/bioinformatics/btp352.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Liberzon, A., Birger, C., Thorvaldsdóttir, H., Ghandi, M., Jill, and Tamayo, P. (2015). The
Molecular Signatures Database Hallmark Gene Set Collection. Cell Systems 1, 417-425.
10.1016/j.cels.2015.12.004.
MacKenzie, A., Hing, B., and Davidson, S. (2013). Exploring the effects of polymorphisms on
cis-regulatory signal transduction response. Trends Mol Med 19, 99-107.
10.1016/j.molmed.2012.11.003.
Mahat, D.B., Kwak, H., Booth, G.T., Jonkers, I.H., Danko, C.G., Patel, R.K., Waters, C.T.,
Munson, K., Core, L.J., and Lis, J.T. (2016). Base-pair-resolution genome-wide mapping of
active RNA polymerases using precision nuclear run-on (PRO-seq). Nature Protocols 11, 1455-
1476. 10.1038/nprot.2016.086.
Martin, M. (2011). Cutadapt removes adapter sequences from high-throughput sequencing reads.
2011 17, 3. 10.14806/ej.17.1.200.
McGinnis, C.S., Patterson, D.M., Winkler, J., Conrad, D.N., Hein, M.Y., Srivastava, V., Hu,
J.L., Murrow, L.M., Weissman, J.S., Werb, Z., et al. (2019). MULTI-seq: sample multiplexing
for single-cell RNA sequencing using lipid-tagged indices. Nature Methods 16, 619-626.
10.1038/s41592-019-0433-8.
Mumbach, M.R., Rubin, A.J., Flynn, R.A., Dai, C., Khavari, P.A., Greenleaf, W.J., and Chang,
H.Y. (2016). HiChIP: efficient and sensitive analysis of protein-directed genome architecture.
Nature methods 13, 919-922.
Murai, J., Zhang, H., Pongor, L., Tang, S.-W., Jo, U., Moribe, F., Ma, Y., Tomita, M., and
Pommier, Y. (2020). Chromatin Remodeling and Immediate Early Gene Activation by SLFN11
in Response to Replication Stress. Cell Reports 30, 4137-4151.e4136.
10.1016/j.celrep.2020.02.117.
Nguyen, A., Yoshida, M., Goodarzi, H., and Tavazoie, S.F. (2016). Highly variable cancer
subpopulations that exhibit enhanced transcriptome variability and metastatic fitness. Nature
communications 7, 1-13.
Nicolas, D., Zoller, B., Suter, D.M., and Naef, F. (2018). Modulation of transcriptional burst
frequency by histone acetylation. Proceedings of the National Academy of Sciences 115, 7153-
7158. 10.1073/pnas.1722330115.
Ostuni, R., Piccolo, V., Barozzi, I., Polletti, S., Termanini, A., Bonifacio, S., Curina, A.,
Prosperini, E., Ghisletti, S., and Natoli, G. (2013). Latent Enhancers Activated by Stimulation in
Differentiated Cells. Cell 152, 157-171. 10.1016/j.cell.2012.12.018.
Parab, L., Pal, S., and Dhar, R. (2022). Transcription factor binding process is the primary driver
of noise in gene expression. PLoS genetics 18, e1010535.
Pedraza, J.M., Garcia, D.A., and Pérez-Ortiz, M.F. (2018). Noise, information and fitness in
changing environments. Frontiers in Physics 6, 83.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Pruitt, K.D., Brown, G.R., Hiatt, S.M., Thibaud-Nissen, F., Astashyn, A., Ermolaeva, O., Farrell,
C.M., Hart, J., Landrum, M.J., McGarvey, K.M., et al. (2013). RefSeq: an update on mammalian
reference sequences. Nucleic Acids Research 42, D756-D763. 10.1093/nar/gkt1114.
Qin, S., Jiang, J., Lu, Y., Nice, E.C., Huang, C., Zhang, J., and He, W. (2020). Emerging role of
tumor cell plasticity in modifying therapeutic response. Signal Transduct Target Ther 5, 228.
10.1038/s41392-020-00313-5.
Raj, A., and Van Oudenaarden, A. (2008). Nature, Nurture, or Chance: Stochastic Gene
Expression and Its Consequences. Cell 135, 216-226. 10.1016/j.cell.2008.09.050.
Raser, J.M., and O'Shea, E.K. (2005). Noise in Gene Expression: Origins, Consequences, and
Control. Science 309, 2010-2013. 10.1126/science.1105891.
Reddy, T.E., Pauli, F., Sprouse, R.O., Neff, N.F., Newberry, K.M., Garabedian, M.J., and Myers,
R.M. (2009). Genomic determination of the glucocorticoid response reveals unexpected
mechanisms of gene regulation. Genome research 19, 2163-2171.
Rodriguez, A.C., Blanchard, Z., Maurer, K.A., and Gertz, J. (2019a). Estrogen Signaling in
Endometrial Cancer: a Key Oncogenic Pathway with Several Open Questions. Hormones and
Cancer 10, 51-63. 10.1007/s12672-019-0358-9.
Rodriguez, J., Ren, G., Day, C.R., Zhao, K., Chow, C.C., and Larson, D.R. (2019b). Intrinsic
Dynamics of a Human Gene Reveal the Basis of Expression Heterogeneity. Cell 176, 213-226
e218. 10.1016/j.cell.2018.11.026.
Schier, A.C., and Taatjes, D.J. (2020). Structure and mechanism of the RNA polymerase II
transcription machinery. Genes & Development 34, 465-488. 10.1101/gad.335679.119.
Schnoes, K.K., Jaffe, I.Z., Iyer, L., Dabreo, A., Aronovitz, M., Newfell, B., Hansen, U., Rosano,
G., and Mendelsohn, M.E. (2008). Research Resource: Rapid Recruitment of Temporally
Distinct Vascular Gene Sets by Estrogen. Molecular Endocrinology 22, 2544-2556.
10.1210/me.2008-0044.
Servant, N., Varoquaux, N., Lajoie, B.R., Viara, E., Chen, C.-J., Vert, J.-P., Heard, E., Dekker,
J., and Barillot, E. (2015). HiC-Pro: an optimized and flexible pipeline for Hi-C data processing.
Genome biology 16, 1-11.
Shaffer, S.M., Dunagin, M.C., Torborg, S.R., Torre, E.A., Emert, B., Krepler, C., Beqiri, M.,
Sproesser, K., Brafford, P.A., Xiao, M., et al. (2017). Rare cell variability and drug-induced
reprogramming as a mode of cancer drug resistance. Nature 546, 431-435. 10.1038/nature22794.
Sheng, M., and Greenberg, M.E. (1990). The regulation and function of c-fos and other
immediate early genes in the nervous system. Neuron 4, 477-485. https://doi.org/10.1016/0896-
6273(90)90106-P.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Shu, S., Lin, C.Y., He, H.H., Witwicki, R.M., Tabassum, D.P., Roberts, J.M., Janiszewska, M.,
Jin Huh, S., Liang, Y., Ryan, J., et al. (2016). Response and resistance to BET bromodomain
inhibitors in triple-negative breast cancer. Nature 529, 413-417. 10.1038/nature16508.
Shvartsur, A., and Bonavida, B. (2014). Trop2 and its overexpression in cancers: regulation and
clinical/ therapeutic implications. Genes & Cancer 6, 84-105. 10.18632/genesandcancer.40.
Sigalova, O.M., Shaeiri, A., Forneris, M., Furlong, E.E., and Zaugg, J.B. (2020). Predictive
features of gene expression variation reveal mechanistic link with differential expression.
Molecular systems biology 16, e9539.
Simeonov, D.R., Gowen, B.G., Boontanrart, M., Roth, T.L., Gagnon, J.D., Mumbach, M.R.,
Satpathy, A.T., Lee, Y., Bray, N.L., Chan, A.Y., et al. (2017). Discovery of stimulation-
responsive immune enhancers with CRISPR activation. Nature 549, 111-115.
10.1038/nature23875.
Single Cell Suspensions from Cultured Cell Lines for Single Cell RNA Sequencing. (2017). 10x
Genomics Document Number CG00054 Rev B.
Stanford, J.L., Szklo, M., and Brinton, L.A. (1986). Estrogen receptors and breast cancer.
Epidemiologic reviews 8, 42-59.
Stuart, T., Butler, A., Hoffman, P., Hafemeister, C., Papalexi, E., Mauck, W.M., Hao, Y.,
Stoeckius, M., Smibert, P., and Satija, R. (2019). Comprehensive Integration of Single-Cell Data.
Cell 177, 1888-1902.e1821. 10.1016/j.cell.2019.05.031.
Suderman, R., Bachman, J.A., Smith, A., Sorger, P.K., and Deeds, E.J. (2017). Fundamental
trade-offs between information flow in single cells and cellular populations. Proceedings of the
National Academy of Sciences 114, 5755-5760. doi:10.1073/pnas.1615660114.
Szustakowski, J.D., Kosinski, P.A., Marrese, C.A., Lee, J.-H., Elliman, S.J., Nirmala, N., and
Kemp, D.M. (2007). Dynamic resolution of functionally related gene sets in response to acute
heat stress. BMC Molecular Biology 8, 46. 10.1186/1471-2199-8-46.
Team, R.C. (2013). R: A language and environment for statistical computing.
Tullai, J.W., Schaffer, M.E., Mullenbrock, S., Sholder, G., Kasif, S., and Cooper, G.M. (2007).
Immediate-Early and Delayed Primary Response Genes Are Distinct in Function and Genomic
Architecture. Journal of Biological Chemistry 282, 23981-23995. 10.1074/jbc.m702044200.
Uhlitz, F., Sieber, A., Wyler, E., Fritsche‐Guenther, R., Meisig, J., Landthaler, M., Klinger, B.,
and Blüthgen, N. (2017). An immediate–late gene expression module decodes ERK signal
duration. Molecular systems biology 13, 928.
Urban, E.A., and Johnston, R.J., Jr. (2018). Buffering and Amplifying Transcriptional Noise
During Cell Fate Specification. Front Genet 9, 591. 10.3389/fgene.2018.00591.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
Uribe, M.L., Marrocco, I., and Yarden, Y. (2021). EGFR in cancer: Signaling mechanisms,
drugs, and acquired resistance. Cancers 13, 2748.
Urrutia, R. (2003). KRAB-containing zinc-finger repressor proteins. Genome Biol 4, 231.
10.1186/gb-2003-4-10-231.
Wang, Z., Zang, C., Rosenfeld, J.A., Schones, D.E., Barski, A., Cuddapah, S., Cui, K., Roh, T.-
Y., Peng, W., Zhang, M.Q., and Zhao, K. (2008). Combinatorial patterns of histone acetylations
and methylations in the human genome. Nature Genetics 40, 897-903. 10.1038/ng.154.
Wei, Y., Gokhale, R.H., Sonnenschein, A., Montgomery, K.M.t., Ingersoll, A., and Arnosti, D.N.
(2016). Complex cis-regulatory landscape of the insulin receptor gene underlies the broad
expression of a central signaling regulator. Development 143, 3591-3603.
Wibisana, J.N., Inaba, T., Shinohara, H., Yumoto, N., Hayashi, T., Umeda, M., Ebisawa, M.,
Nikaido, I., Sako, Y., and Okada, M. (2022). Enhanced transcriptional heterogeneity mediated by
NF-κB super-enhancers. PLOS Genetics 18, e1010235. 10.1371/journal.pgen.1010235.
Wollman, R. (2018). Robustness, Accuracy, and Cell State Heterogeneity in Biological Systems.
Curr Opin Syst Biol 8, 46-50. 10.1016/j.coisb.2017.11.009.
Wood, S.N. (2006). Generalized additive models: an introduction with R (chapman and
hall/CRC).
Wu, S., Li, K., Li, Y., Zhao, T., Li, T., Yang, Y.-F., and Qian, W. (2017). Independent regulation
of gene expression level and noise by histone modifications. PLOS Computational Biology 13,
e1005585. 10.1371/journal.pcbi.1005585.
Zhang, G., Zhao, Y., Liu, Y., Kao, L.-P., Wang, X., Skerry, B., and Li, Z. (2016). FOXA1
defines cancer cell specificity. Science Advances 2, e1501473. doi:10.1126/sciadv.1501473.
Zhang, J., Lee, D., Dhiman, V., Jiang, P., Xu, J., McGillivray, P., Yang, H., Liu, J., Meyerson,
W., Clarke, D., et al. (2020). An integrative ENCODE resource for cancer genomics. Nature
Communications 11. 10.1038/s41467-020-14743-w.
Zhang, J., Zhu, W., Wang, Q., Gu, J., Huang, L.F., and Sun, X. (2019). Differential regulatory
network-based quantification and prioritization of key genes underlying cancer drug resistance
based on time-course RNA-seq data. PLoS computational biology 15, e1007435.
Zhang, Y., Liu, T., Meyer, C.A., Eeckhoute, J., Johnson, D.S., Bernstein, B.E., Nusbaum, C.,
Myers, R.M., Brown, M., and Li, W. (2008). Model-based analysis of ChIP-Seq (MACS).
Genome biology 9, 1-9.
.CC-BY 4.0 International licenseavailable under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprintthis version posted March 15, 2023. ; https://doi.org/10.1101/2023.03.14.532457doi: bioRxiv preprint
