Molecular Cell

Review

What’s Luck Got to Do with It: Single Cells,Multiple Fates, and Biological Nondeterminism

Orsolya Symmons1,* and Arjun Raj1,*1Department of Bioengineering, University of Pennsylvania, Philadelphia, PA 19104, USA*Correspondence: usymmons@yahoo.com (O.S.), arjunrajlab@gmail.com (A.R.)http://dx.doi.org/10.1016/j.molcel.2016.05.023

The field of single-cell biology has morphed from a philosophical digression at its inception, to a playgroundfor quantitative biologists, to a major area of biomedical research. The last several years have witnessed anexplosion of new technologies, allowing us to apply evenmore of themodernmolecular biology toolkit to sin-gle cells. Conceptual progress, however, has been comparatively slow. Here, we provide a framework forclassifying both the origins of the differences between individual cells and the consequences of those differ-ences. We discuss how the concept of ‘‘random’’ differences is context dependent, and propose thatrigorous definitions of inputs and outputs may bring clarity to the discussion. We also categorize ways inwhich probabilistic behavior may influence cellular function, highlighting studies that point to exciting futuredirections in the field.

IntroductionThere is a fairly broad appreciation in the biological community

that the behavior of individual cells may vary from the popula-

tion average, giving rise to an entire field of ‘‘single-cell

biology’’ (Balazsi et al., 2011; Eldar and Elowitz, 2010; Raj

and van Oudenaarden, 2008; Trapnell, 2015). Early, pioneering

work in bacteria (Benzer, 1953; Novick and Weiner, 1957) and

mammalian cells (Ko et al., 1990) provided convincing demon-

strations that cell-to-cell variability is indeed a fact of life. These

studies are all the more remarkable given the limited experi-

mental tools available at the time, which often required making

inferences based on clever experimental design and specifics

of the system in question. Fast forward a few decades, and

we can make measurements in single cells those researchers

probably could not even have dreamed of—live cell imaging

of transcription with single-molecule resolution, measuring

the entire transcriptome of thousands of single cells, and who

knows what will be possible 2 years from now. Yet while those

early years were marked with considerable theoretical discus-

sion of the basis and consequences of the life of a single cell

(Arkin et al., 1998; Peccoud and Ycart, 1995; Schrodinger,

1944; Spudich and Koshland, 1976), such discussions have

fallen by the wayside as our drive for quantification has far out-

paced our justification for making those measurements to

begin with. As single-cell biology has recently joined the trend

toward industrialization that is sweeping through molecular

biology in general, we feel the time is ripe for returning to

some of these fundamental questions before we embark on

massive data-gathering exercises. Here, our goal is to discuss

a potential framework for classifying studies of single-cell

biology.

Where to begin with such a framework? We think it instructive

to consider thatmany biologists, especially those of the develop-

mental variety, might be forgiven for saying, ‘‘Single-cell

biology? Isn’t that what we’ve been calling ‘biology’ for de-

cades’’? Certainly, the fact that individual cells, different tissue

types, and even multicellular organisms can do different things

788 Molecular Cell 62, June 2, 2016 ª 2016 Elsevier Inc.

with the same genome is hardly news (see examples in Figure 1).

A potential starting point for a more useful discussion is to

develop a conceptual classification of ways in which we think

about differences between single cells. To make this concrete,

let’s consider a side-by-side comparison between two cells.

Much of the focus of the field has been on listing how these

two cells may be different at the molecular level. In particular,

our tools now enable us to measure the differences in the molec-

ular state of a cell with extraordinary breadth and accuracy

(though perhaps not both simultaneously), as has been reviewed

thoroughly elsewhere (Itzkovitz and van Oudenaarden, 2011; Raj

and van Oudenaarden, 2009; Trapnell, 2015).

Perhaps more interesting is to consider why these two cells

are different. Broadly, there are two non-exclusive reasons

why two cells could be different from each other: deterministic,

in which the cells receive different instructions, leading to

different outcomes, and probabilistic, in which cells receiving

the same instructions can have different outcomes. The latter

has often been termed ‘‘random’’ or ‘‘stochastic,’’ and we

believe that this has led to some conceptual difficulties, arising

from inconsistencies in defining what it means to truly be

random. For any given difference between cells, say, the expres-

sion of a particular gene, one must assume that difference

arises—deterministically—from some other difference between

the cells in the past, say, the spatial configuration of transcription

factors regulating the gene. This ‘‘cellular butterfly effect’’

makes the very word ‘‘random’’ difficult to rigorously justify.

We will later propose instead a framework centered around

more concrete operational definitions based on mappings be-

tween inputs and outputs (Figure 2). For example, inputs could

be defined as the abundances of all transcriptional regulators,

the output could be defined as the abundance of the target

gene’s mRNA, and the mapping would then be the process of

transcription itself. A deterministic mapping is one in which the

values of a particular input will always give the same output.

A probabilistic mapping is one in which a given set of inputs

yields different outputs—in the example of gene expression,

A

ON

OFF

genotype organismalphenotype

genotype organismalphenotype

B

C

fertilized oocyte

organismalphenotype

genotype

?

?

?

?

genetically identicaldaughter cells

quadruplets, with different shield patterning

black fur

orange fur

hetero-chromatin

random X inactivation and transcription

Figure 1. Variable Phenotypic Interpretations of Genomic Information(A) Position-effect variegation gives rise to red and white patches in the Drosophila eye. The phenotype is due to imperfect spreading of pericentric hetero-chromatin; the white gene (red box) is silenced in some cells (white patches), but remains expressed in others (red patches). Copyright Cold Spring HarborLaboratory Press (Elgin and Reuter, 2013).(B) Calico cats have patches on differently colored fur (in this case black and orange, on a white background). The animals are heterozygous for a gene, with oneallele causing the orange tabby and the other the black color. Random inactivation of either one or the other copy of the X chromosome in individual cells givescalico cats their characteristic patches of color. Picture of calico cat by Lisa Ann Yount, https://flic.kr/p/CZ7sQm (CC0 1.0).(C) Nine-banded armadillos are a polyembryonic species, with a single fertilized egg typically giving rise to quadruplets. While these genetically identical in-dividuals seem very similar at first glance, some traits, such as the patterning of the head shield (h, with two different patterns shown in the red circles) and thebanded shield (b) can be highly variable. Image of armadillos from Vogt (2015), with permission of Springer. Photo: Brian Bagatto.

Molecular Cell

Review

different abundances of the target gene’smRNA.Wewill discuss

various mappings in biology and how various submappings

within those mappings reveal the ways in which probabilistic

behavior is generated and controlled.

With this framework, the question of what differences actually

matter for the cell becomes easier to conceptualize. Here, one

can define the output of a mapping to be, say, a phenotype of

interest. For a given set of inputs, does one get a probabilistic

output? In that case, such a mapping would be considered a

diversity-generating mapping (Figure 2A). For example, if we

consider the mapping between genotype and phenotype, a

probabilistic phenotype would be an example of a diversity-

generating map. Conversely, a mapping that takes multiple in-

puts to a single reliable output would be considered a buffering

Molecular Cell 62, June 2, 2016 789

A

B

C

Figure 2. Mappings of Inputs to Outputs Can Either Be Deterministically Specified, Diversity Generating, or Buffering(A) Illustration of the three types of mappings.(B) Illustration of hidden variables that explain a seemingly probabilistic mapping, with an example of cell size as a hidden variable (from Padovan-Merhar et al.,2015).(C) Diagram of layers of specification in the retinal mosaic in Drosophila melanogaster. Reprinted by permission from Macmillan Publishers Ltd: Nature (Wernetet al., 2006), copyright 2006. Individual ommatidia contain R7 and R8 cells, which either express Rh3 and Rh5 (pale ommatidia), or Rh4 and Rh6, respectively(yellow ommatidia). The specification results from stochastic expression of the spineless transcription factor. Ultimately, the organismal phenotype is�30% paleversus �70% yellow ommatidia.

Molecular Cell

Review

mapping (Figure 2A). For example, consider induced pluripotent

stem cells: our increasingly comprehensive molecular assays

have revealed that they often differ remarkably from the embry-

790 Molecular Cell 62, June 2, 2016

onic stem cells they are meant to mimic, but the bottom line is

that you can use these cells to generate a full mouse (Takahashi

and Yamanaka, 2006).

Molecular Cell

Review

Wewill highlight a number of examples that point to interesting

new types of mappings, suggesting that many biological pro-

cesses consist of layers of both determinism and nondetermin-

ism. We hope that viewing single-cell biology through this lens

may reinvigorate debates around the concepts at the core of

this still rapidly evolving field.

How Cells Are Different from Each Other: Let Me Countthe WaysResearchers have known for decades that genetically identical

cells can differ from each other: as early as 1961, Mary Lyon

(Lyon, 1961) reported mosaic inactivation of the X chromosome

in female mice. In 1989, using the then newly developed tech-

nique of PCR, scientists showed the presence of extremely lowly

expressed tissue-specific genes in bulk RNA from other tissues,

which they inferred to be ‘‘illegitimate’’ transcripts that were pre-

sent in only one or few cells (Chelly et al., 1989). Further refine-

ment of molecular techniques has unearthed a new world of

cellular diversity, and over the last several years, researchers

have developed a plethora of new tools that have revealed

how cells are different from each other with an unprecedented

level of depth, breadth, and variety. A number of excellent re-

views discussing these tools and their findings already exist

(Crosetto et al., 2015; Itzkovitz and van Oudenaarden, 2011),

and therefore we only provide an overview here to give context

for our further discussion.

One of the first areas in which people measured cell-to-cell

variability was in gene expression, by measuring the levels of

mRNA and proteins. Initially, researchers measured expression

levels by simply quantifying the amount of GFP in the cell via fluo-

rescence microscopy or flow cytometry (Elowitz et al., 2002; Oz-

budak et al., 2002; Raser and O’Shea, 2004). Rapidly, these

techniques evolved to measure variability in GFP levels across

panels of genes (Bar-Even et al., 2006) or even across all genes

(Newman et al., 2006), revealing, among other things, that

stress-response genes tended to be more variable than house-

keeping genes.

In parallel, the development of single-molecule microscopy

techniques enabled new science-fiction-like levels of accuracy,

enabling us to measure cellular differences with incredible quan-

titative precision by counting individual molecules in single cells.

One thread has evolved around the visualization of individual

mRNAs in single cells, both in fixed (Femino et al., 1998; Levsky

et al., 2002; Raj et al., 2006, 2008; Zenklusen et al., 2008) and

living cells (Bertrand et al., 1998; Chubb et al., 2006; Coulon

et al., 2014; Golding et al., 2005; Larson et al., 2011). Over the

years, these methods have matured in terms of the types of

processes that one can measure, now including sensitive mea-

surements of the transcriptional process itself (Levesque and

Raj, 2013; Levsky et al., 2002; Little et al., 2013), nuclear export

(Grunwald and Singer, 2010), and allele-specific expression

(Ginart et al., 2016; Hansen and van Oudenaarden, 2013; Lev-

esque et al., 2013). New probing strategies have also greatly

increased the breadth of measurements, enabling on the one

hand the measurement of thousands of genes, one at a time,

across very large numbers of cells (Battich et al., 2013), and

on the other hand allowing measurements of transcript abun-

dance of tens to hundreds or even thousands of genes simulta-

neously in individual cells (Chen et al., 2015; Levesque and Raj,

2013; Lubeck et al., 2014; Lubeck and Cai, 2012), with direct

in situ sequencing on the horizon (Ke et al., 2013; Lee et al.,

2014).

Single-protein measurements present a greater technical

challenge, but clever combinations of different technologies

have allowed the absolute quantification of protein numbers,

pioneered in large part by the Xie and Xiao labs (Cai et al.,

2006; Choi et al., 2008; Hensel et al., 2012; Yu et al., 2006),

even parallelized to measure thousands of genes in E. coli (Tani-

guchi et al., 2010). Techniques such as mass cytometry enable

the detection of protein levels by flow cytometry for dozens of

proteins at a time, allowing also for the detection of protein mod-

ifications that are essential parts of cell signaling (Bendall et al.,

2012). Spatial localization of signaling molecules in single cells,

such as nuclear translocation (Cai et al., 2008; Tay et al.,

2010), can also reveal signaling behavior in single cells.

While most of the examples cited above were mostly limited

to the study of one or a few genes at a time through micro-

scopy, much of the recent interest in single-cell biology has

been driven by the massive increase in the scale of measure-

ments, primarily advances in sequencing. Single-cell RNA-

sequencing has allowed us to probe the entire transcriptome

of individual cells, and while technical questions about accu-

racy remain (Grun et al., 2014; Marinov et al., 2014; Wu et al.,

2014), the ability to measure even coarse transcriptomes

from large numbers of cells (Klein et al., 2015; Macosko

et al., 2015) is poised to transform the sophistication of the

questions we can ask. Developments in mass spectrometry

suggest that single-cell proteomics (Bjornson et al., 2013) and

metabolomics (Zenobi, 2013) may be on the verge of a

revolution akin to that of the RNA field. Analysis of the protein

content of single cells is likely to provide us with a far richer

and more comprehensive picture of cellular variability in the

near future.

Despite our now rather mature ability to measure RNA and

protein at the single-cell level, a host of open questions remains.

It will be increasingly important to integrate different types of

measurements and simultaneously monitor regulatory activity,

RNA and protein levels, as well as the metabolic and signaling

state of individual cell, to fully capture the relationship between

these processes and their relative importance for biological func-

tion. For example, simultaneous measurements of RNA and pro-

tein levels in single cells (Albayrak et al., 2016; Raj et al., 2006;

Taniguchi et al., 2010) have shown that depending on half-lives,

mRNA, and protein levels may or may not correlate strongly be-

tween cells, and new tools have even enabled the visualization of

the initial rounds of translation (Halstead et al., 2015; Katz et al.,

2016; Wu et al., 2015). At the same time, recent work has also at-

tempted to link transcription factor localization and binding to

transcription. Using cutting edge microscopy (Elf et al., 2007;

Hammar et al., 2012) and probing technology (Shah and Tyagi,

2013), efforts to relate the binding of transcription factors to tran-

scription itself are just now shedding light on these relationships

(Larson et al., 2011; Sepulveda et al., 2016; Shah and Tyagi,

2013; Xu et al., 2015), and single-cell reporters of methylation

are on the horizon as well (Stelzer et al., 2015). The advent of

genome-wide techniques for measuring transcription factor

Molecular Cell 62, June 2, 2016 791

Molecular Cell

Review

binding provides a glimpse of what the future holds in this regard

(Buenrostro et al., 2015; Cusanovich et al., 2015).

The integration of spatial information to understand the rela-

tionship between cells in space and time is another frontier in

the field that will be critical for linking single-cell phenomena to

organismal behavior. Imaging studies are a natural starting point,

and RNA hybridization studies have already begun to assess the

degree of heterogeneity directly in tissue (Bahar Halpern et al.,

2015a, 2015b; Bahar Halpern and Itzkovitz, 2016). Exciting

new developments augmenting sequencing-based transcrip-

tomics with spatial information have led to the development of

spatial transcriptomics (Achim et al., 2015; Junker et al., 2014;

Satija et al., 2015), and as mentioned, in situ sequencing may ul-

timately provide an even more direct spatial picture of expres-

sion (Lee et al., 2014). Spatial information can provide not just

context, but by proxy, information on cell lineage, showing

that, for instance, expression profiles of closely related (e.g., sis-

ter) cells tend to be more correlated (Bai et al., 2010; Cote et al.,

2016; Duffy et al., 2012). Furthermore, combining time-lapse mi-

croscopy with subsequent transcriptional analysis can explicitly

demonstrate how cells inherit expression patterns through cell

divisions (Cote et al., 2016).

Why Are Cells Different from Each Other?As the single-cell toolkit expands, the question of how to parse

all these data remains largely unanswered, however. Even if

we could count every molecule in the cell, would that necessarily

lead to more understanding of the biology of single cells? Just as

Tycho Brahe’s comprehensive star charts were not the key to

understanding the motion of heavenly bodies, we think that

more careful conceptual thought must go into understanding

exactly why cells are different from each other; otherwise, we

will continue with quantification sans justification ad infinitum

(Mellis and Raj, 2015).

Gene expression provides an excellent case study to examine

the reasons why cells are different: the landmark studies of Elo-

witz et al. and Ozbudak et al. (Elowitz et al., 2002; Ozbudak et al.,

2002) clearly demonstrated using fluorescent proteins that

expression levels can vary between otherwise identical-seeming

cells. These papers in many ways marked the beginning of the

current era of quantitative single-cell biology.

The variability identified by these studies raised the funda-

mental question: why are these cells different from each other?

A priori, there are a number of possibilities. Perhaps the most

aesthetically pleasing is the chemical one, in which the small

numbers of molecules and their random collisions—in this

case, say, RNA polymerases with DNA and ribosomes with

RNA (Arkin et al., 1998; Kepler and Elston, 2001; McAdams

and Arkin, 1997; Paulsson, 2005; Thattai and van Oudenaarden,

2001)—lead to variability in transcription and translation. Yet im-

plicit in this explanation is the notion that each cell is otherwise

completely equivalent to the next one. What if there were other

‘‘hidden variables’’ explaining why the expression of the cells

differed, i.e., if not all inputs to the mapping had been fully

defined (Figure 2B)? The beauty of the Elowitz experiment was

that it not only revealed cell-to-cell variability, but that it also pro-

vided an explicit demonstration of a probabilistic mapping that

takes hidden variables into account. Experimentally, they con-

792 Molecular Cell 62, June 2, 2016

structed two almost identical but experimentally distinguishable

copies of a gene in the cell, and any deviations between these

two genes were interpreted as ‘‘intrinsically random’’ variability,

whereas deviations shared between the two genes, but different

in one cell and another cell, were deemed ‘‘extrinsic’’ variability,

i.e., variability due to hidden or not-considered factors. Elowitz

et al. found that both types of variability exist.

The demonstration that hidden factors exist provides an

imperative to more carefully define the inputs to the gene

expression mapping. Indeed, consider the case of lambda

phage: seminal theoretical work demonstrated that pure chem-

ical noise could generate the divergent cell fates of lysis and

lysogeny (Arkin et al., 1998), but subsequent experimental

work has shown that much of this diversity is actually the result

of variability in ‘‘hidden factors’’ (St-Pierre and Endy, 2008;

Zeng et al., 2010), such as cell size, position in cell cycle, and

even the subcellular localization of virus infection. Other studies

have revealed a host of hidden factors in other systems as well.

For example, in mammalian cells, even the simple fact that

larger cells have more RNA in them overall (Kempe et al.,

2015; Padovan-Merhar et al., 2015) shows that many hidden

variables may lay in plain sight. (Importantly, Padovan-Merhar

et al. followed up with experiments to explicitly demonstrate

that this results from a global volume-dependent transcriptional

control mechanism to maintain transcript concentration rather

than just count.) Additional factors influencing variability in

expression include cell cycle (Buettner et al., 2015; Zopf et al.,

2013), mitochondrial variability (Guantes et al., 2015; Johnston

et al., 2012), heterogeneity in cell culture medium (Guo et al.,

2016), temperature (Arnaud et al., 2015), and a plethora of other

factors (Battich et al., 2015).

It is tempting, then, to imagine that a proper determination of

intrinsic, random variability would require factoring out all other

hidden factors, and indeed many studies try to ‘‘control’’ for as

many of these factors as possible. Ultimately, though, this quix-

otic endeavor would leave us empty handed, because factoring

out all possible hidden variables would also remove all random-

ness. After all, setting aside quantummechanical considerations

for now, if we could measure the locations and trajectories of

every molecule, and could then (with sufficient simulation power)

predict why one copy of Elowitz’s gene produced more fluores-

cent protein than the other, would the differences then be

considered truly random? Although these philosophical musings

are worth debating, the point is that there are always more hid-

den factors, and so the motivation for the mapping concept we

have introduced is to sidestep these discussions and allow for

a framework to practically discuss the origins of cellular vari-

ability. In particular, the concept of submappings that reflect

both our increased ability to measure various inputs and outputs

and our theoretical considerations of what layers of abstraction

aremost important, provides a practical means by which to clas-

sify the origins of cellular variability. For instance, now that we are

able tomeasure transcription factor concentration and transcrip-

tion directly in single cells (Sepulveda et al., 2016), we can

decompose the original Elowitz mapping (from cellular concen-

trations of all potential regulators to fluorescent protein levels)

into two submappings: first, from cellular concentrations of a

particular regulator to transcription, and second, from transcript

Molecular Cell

Review

abundance to protein abundance. Being explicit about such de-

compositions (and the definitions of inputs and outputs) allows

us to precisely describe what the sources of probabilistic

behavior are.

Probabilistic Gene Expression

Gene expression provides perhaps the best-studied set of ex-

amples of probabilistic mappings and submappings. These sub-

mappings involve a plethora of molecular processes, including

transcription, translation, RNA degradation, binomial partitioning

upon cell division (Rosenfeld et al., 2005), alternative splicing

(Waks et al., 2011), and nuclear trafficking (Bahar Halpern

et al., 2015a; Battich et al., 2015). Yet for most genes, the root

source of probabilistic expression levels is most likely the low

copy number of the gene itself, combined with the fact that in

many instances, the transcription of genes is itself pulsatile,

with transcription occurring in ‘‘bursts’’ (Bahar Halpern et al.,

2015b; Chubb et al., 2006; Golding et al., 2005; Levsky et al.,

2002; Raj et al., 2006; Suter et al., 2011). How do we know these

bursts are a manifestation of the probabilistic execution of the

transcriptional mapping? To the extent that we believe the intra-

cellular milieu is the input to the mapping, then an Elowitz-style

experiment would reveal probabilistic behavior. In mammalian

cells, imaging studies have revealed that bursts from the two al-

leles are largely uncoordinated (Levesque and Raj, 2013), thus

showing that bursts themselves are a major source of nondeter-

minism in gene expression. Similarly, multiple independent

studies that used gene expression measurements from a collec-

tion of cell lines where a transgene was positioned at different

genomic locations have shown that the difference in gene

expression variance is rooted in the locus-dependent pattern

of transcriptional bursts (Dey et al., 2015; Singh et al., 2010; Vi-

nuelas et al., 2013).

A major remaining gap is to disentangle the submappings be-

tween the transcriptional regulators and the bursts themselves,

i.e., between the biochemical underpinnings of transcription

and the phenomenology of bursts. In the case of bacteria, the

work of Chong et al. has shown that DNA supercoiling can result

in bursts of transcription (Chong et al., 2014), while other work

measuring binding of transcription factors simultaneously with

transcription revealed that transcription factor occupancy was

too rapid to lead to bursts in and of itself (Sepulveda et al.,

2016) (Jones et al., 2014). In eukaryotes, and particularly higher

eukaryotes, the situation is considerably murkier. The most

intuitive candidate for bursting is the binding of a transcription

factor; however, the timescales of transcription factor binding

and unbinding are considerably faster than that of transcrip-

tional bursts, and most strikingly, variability in transcription fac-

tor concentration and binding does not seem to propagate to

the level of transcriptional bursts (Shah and Tyagi, 2013; Xu

et al., 2015). Thus, the mapping between transcription factor

milieu and instantaneous transcriptional activation seems highly

probabilistic.

Looking elsewhere, evidence supports some role for nucleo-

some positioning (Brown et al., 2013; Raser and O’Shea,

2004), promoter structure (Dadiani et al., 2013; Suter et al.,

2011), and transcription factor concentration (Octavio et al.,

2009; Senecal et al., 2014) in regulating transcriptional bursting.

In addition, numerous studies have pointed to chromatin modi-

fiers, particularly histone deacetylases, as modulators of

bursting behavior (Batenchuk et al., 2011; Dar et al., 2012; Raj

et al., 2010; Suter et al., 2011; Weinberger et al., 2012), although

the pleiotropic nature of these perturbations makes the case

more suggestive than definitive. Another possibility is that

spatial organization of nuclear chromatin may influence bursting,

either by providing a scaffold for stochastic interactions between

cis-regulatory elements and proteins, or through the probabi-

listic interaction between regulatory elements and promoters

(Hacisuleyman et al., 2014; Maamar et al., 2013; Misteli, 2007;

Schoenfelder et al., 2010; Splinter and de Laat, 2011). While

simultaneous measurement of three-dimensional organization,

transcription factor occupancy, and bursting at the single-cell

level is not yet possible due to technical limitations, indirect

data already support links between individual elements of this

hypothesis (Amano et al., 2009; Liu et al., 2014; Noordermeer

et al., 2011), although we are not sure how general these findings

may be.

It is important to note that transcriptional bursts are not purely

nondeterministic—to the extent that transcription is itself regu-

lated, mapping must in fact work through bursts in some way.

Recent work has shown that the two ‘‘dials’’ provided by tran-

scriptional bursts (i.e., transcriptional burst size and frequency)

can be tuned independently, and may represent different ways

to tune gene expression (Batenchuk et al., 2011; Dar et al.,

2012; Octavio et al., 2009; Padovan-Merhar et al., 2015; Raj

et al., 2006; Senecal et al., 2014; Suter et al., 2011; Weinberger

et al., 2012).

While no overarching rules for regulation of transcriptional

bursts have emerged, our suspicion is that trans factors largely

regulate transcriptional burst size and cis regulatory factors

largely regulate transcriptional burst frequency (and that much

of the confusion arises from the fact that the line between cis

and trans regulation can be a bit blurry). Circumstantially, it

seems that burst frequency varies depending on chromosomal

context (Becskei et al., 2005; Raj et al., 2006). In a more direct

example, a recent study from our lab has demonstrated that

changes to looping of the enhancer to the promoter, when

appropriately isolated from other regulatory effects, affect burst

frequency specifically (Bartman et al., 2016). At the same time,

phenomenological studies show that the cell may turn specific

dials to solve particular biological problems, such as dosage

compensation during DNA replication (Padovan-Merhar et al.,

2015; Skinner et al., 2016; Yunger et al., 2010) or dosage

compensation for cell size (Padovan-Merhar et al., 2015). Eluci-

dating the mappings between mechanisms of gene regulation,

transcriptional bursts, and consequences of this random firing

remains a major challenge in the field, and will ultimately be crit-

ical to understanding this important source of probabilistic

behavior in gene expression.

Probabilistic Cellular Identity

In the case of gene expression, the framework of deterministic

versus probabilistic maps looks rather similar to the experi-

ments from Elowitz et al. Conceptually, however, the mapping

framework can encompass many other processes and aspects

of single-cell biology. Consider cell-type determination in multi-

cellular organisms, a topic that has gained renewed excitement

after single-cell RNA-sequencing experiments revealed the

Molecular Cell 62, June 2, 2016 793

Molecular Cell

Review

presence of new cell types and cell states, including new inter-

mediates (Trapnell, 2015). Let’s say we measure the transcrip-

tome profiles of two cells and find a set of genes with differential

expression. Often, these differences are assumed to be part of

a cell-type specification program, and many algorithms to

reconstruct cell differentiation trajectories make this assump-

tion implicitly (Bendall et al., 2014; Trapnell, 2015; Trapnell

et al., 2014). Yet it is possible that many of the differences be-

tween these cells are probabilistic rather than deterministic.

As we will discuss, rigorous demonstration of this fact is diffi-

cult, and many seemingly probabilistic mappings become

more deterministic as hidden variables are exposed, but we

will also highlight examples in which we think the case for prob-

abilistic mappings is strong, describing how the definition of in-

puts allows one to make that inference. We also point to results

that hint that the relative homogeneity we typically observe

may mask hidden underlying probabilistic behavior, suggesting

that cell-fate specification may be composed of layers of diver-

sity generating and subsequent buffering mappings. A major

challenge for the field will be to settle on the appropriate

levels of abstraction for these mappings to enable practical

application.

As an illustration of the difficulties associated with definitively

showing that cellular behavior is probabilistic as opposed to

deterministic, consider the case of embryonic stem cells.

When grown in culture, these cells often exhibit variable levels

of expression of key regulators such as Nanog. The question

is, what is the status of the cells with low levels of Nanog? Are

they intermediate states of cells on their (deterministic) way to

differentiation? Or, given the signals and other inputs they

receive, are they in a probabilistic, transient (and hence revers-

ible) ‘‘primed’’ state in which they are exploring several potential

lineages (Abranches et al., 2014; Kalmar et al., 2009; Singer

et al., 2014)? One means by which to argue for transient (and

thus potentially probabilistic) intermediates is to sort out high

and low Nanog populations of cells and then see if these

subpopulations will eventually revert to the original population’s

distribution; similar experiments have also been done in cancer

(Gupta et al., 2011) and hematopoiesis (Chang et al., 2008).

On the face of it, such experiments confirm state reversion

and thus a probabilistic process, though careful analyses of

population dynamics both in stem cells (Nair et al., 2015) and

hematopoiesis (Pina et al., 2012) raise the possibility that such

effects are mostly due to differential growth dynamics. At the

same time, observations of variability in Nanog expression

levels in vivo raise the possibility that such variability is not

just an artifact of cell culture (Smith, 2013). The difficulty of con-

firming probabilistic behavior in vivo, however, is verifying that

the inputs to the mapping (e.g., signaling pathways, regulatory

milieu) are indeed constant and not predetermined by some

developmental pathway. The ongoing controversy serves to

highlight some of the difficulties in establishing whether behavior

is truly probabilistic. Such studies could have immense practical

implications: one major barrier to the therapeutic use of stem

cells is their seemingly probabilistic tendency to differentiate

into multiple different lineages, often resulting in poor perfor-

mance (Cote et al., 2016). If we can specify a layer of mapping

at which the lineage chosen is deterministic, then that layer

794 Molecular Cell 62, June 2, 2016

may hold the key to directing stem cells toward more homoge-

neous fates.

As researchers have delved deeper into regulatory mecha-

nisms, other examples of probabilistic behavior have been re-

vealed to be far more deterministic than initially thought, much

as was the case for lambda phage. For instance, in mouse blas-

tocyst development, work on gene expression and cell fate

decisions initially suggested that early genes have a seemingly

stochastic gene expression pattern, which later stabilizes and

defines cell lineages through signal reinforcement (Dietrich and

Hiiragi, 2007; Ohnishi et al., 2014; Plusa et al., 2008). This is in

contrast with new data, which suggest not only that clear differ-

ences can already be identified between cells (Biase et al., 2014),

but that Sox2 (and Oct4) activity and binding predict cell fate

(Goolam et al., 2016; White et al., 2016) much earlier than previ-

ously suggested. Another example of a hidden variable is that of

mice that can have either five or six lumbar vertebrae. While

seemingly a probabilistic choice, a clever set of embryo transfer

experiments showed that the maternal environment can influ-

ence the seemingly random decision of the progeny (McLaren

and Michie, 1958).

Stem cell division provides us with another example of poten-

tially probabilistic behavior, this time the maintenance of tissue-

specific stem cells. Tissue-specific stem cells are responsible for

maintaining and replacing adult tissues, but the correct balance

between maintaining the stem cell population and differentiation

of cells is critical for tissue function. Recent work (reviewed by

Krieger and Simons, 2015) has shown that in addition to the

classic model of differentiation, wherein this balance is achieved

through asymmetric cell divisions, themaintenance of adult stem

cell can also be determined at the population level, with proba-

bilistic loss and replacement of individual stem cells. Thus, while

inputs exist that can deterministically dictate the fate of daughter

cells in some cases (Inaba and Yamashita, 2012), experiments

using time-lapse microscopy, lineage tracing, and modeling

have shown that in other situations the stem cell population is

dynamically being turned over, with individual clones constantly

being lost and replaced by others (Doupe et al., 2012). However,

while this behavior is probabilistic at the population level, it is still

unclear precisely what, if any, cues influence these changes in

cell state, so the details of the mappings from one cell’s state

to the states of its progeny (and back) are fertile ground for

further study.

Collectively, these studies show that the concept of probabi-

listic mapping between inputs and outputs likely occurs in a

wide variety of single-cell contexts and layers of specification,

and discriminating between deterministic and probabilistic

behavior could have profound consequences. We have here

chosen just a few of the many studies that illustrate the concep-

tual difficulties of showing a particular phenomenon is truly

random. As our tools become more sophisticated, we can now

often find the ‘‘cause’’ for a particular probabilistic-seeming

event. The question that these studies raise is which of these

previously hidden variables are practically relevant and which

can be effectively ignored. By making those choices explicit,

the mapping framework provides a way to avoid the ‘‘kick the

can down the road’’ issue that plagues the interpretations of

many such studies.

Molecular Cell

Review

Plasticity in Patterning and Other Organismal Traits

Are there any probabilistic behaviors at practical levels of

description at all, then? Consider one of the most broad map-

pings, namely from genotype to organismal phenotype. Depend-

ing on our definitions of inputs and outputs, this specification

program would seem rather deterministic indeed: identical

twins, inbred animal lines, or genetic clones provide a striking

demonstration of this fact, with genetic equality often leading

to virtually indistinguishable physical characteristics. This is of

course not to discount the many examples of phenotypic diver-

sity in genetically identical animals (extensively reviewed in Vogt,

2015), including humans (Zwijnenburg et al., 2010), but in broad

strokes, this mapping seems highly deterministic. Yet this

apparent homogeneity may mask layers of diversity generating

and subsequent buffering mappings. As an example, consider

the developmental lineage. While C. elegans is famous for its

almost completely stereotypic lineage, most organisms have

far more plastic lineages. As such, the precise lineage itself is

not deterministic but rather likely to be probabilistic, although

the ultimate ‘‘phenotype’’ of the organism may buffer this vari-

ability into a similar-seeming animal. Potentially, there are

many layers of diversity generation and subsequent buffering.

The careful definition of inputs and outputs to submappings

can make such layering effects more clear. Consider, for

example, photoreceptor selection in the developing Drosophila

eye (Figure 2C) (Wernet et al., 2006). Here, the fate of a particular

ommatidia as being either ‘‘pale’’ or ‘‘yellow’’ light responsive de-

pends on the expression of a particular transcription factor

(spineless). Some ommatidia have cells with high levels of spine-

less and others do not, thus generating diversity in ommatidia

fate. This example also provides a useful means of establishing

probabilistic evaluation of a well-defined set of inputs—in an

array of otherwise identical ommatidia, the pale ommatidia are

randomly interspersed among the yellow ones, and such config-

urational randomness strongly indicates a probabilistic evalua-

tion. It is worth mentioning, as well, that at a broader phenotypic

level, the eye produces roughly the same proportion of pale and

yellow ommatidia irrespective of the precise spatial configura-

tion; thus, at that level, the functional differences between organ-

isms are low, providing an example of masked variability.

Diversity generation and subsequent buffering may in fact

prove to be the rule rather than the exception. Indeed, there is

mounting evidence that at least some elements of embryonic

patterning, while deterministic in outcome, do not require deter-

ministic transcriptional and cellular inputs, but instead buffer

variability through signaling and self-organization. Much of

embryology is focused on understanding the processes underly-

ing this developmental robustness, and some excellent reviews

exist on this topic (Felix and Barkoulas, 2015; Martinez Arias

et al., 2013; Umulis et al., 2008). We will therefore only highlight

one case: the coordinated oscillations in gene expression that

occur during vertebrate segmentation and lead to the periodic

formation of new segments. These oscillations occur robustly

and coordinately within embryos, but become unstable and

noisy when single cells are isolated from the embryo (Masamizu

et al., 2006). Oscillations can be synchronized through Delta-

Notch signaling (Jiang et al., 2000), and in a recent series of ex-

periments Tsiairis and Aulehla showed that (Tsiairis and Aulehla,

2016) coordinated behavior can be replicated in vitro, when cells

were dissociated and cultured together in a dish. Moreover, by

mixing cells with different oscillation frequencies they found

that they could reestablish synchronous oscillations with a new

frequency that corresponded to the average of the input fre-

quencies. These results serve as an excellent example of buff-

ering probabilistic behavior. It will be very important to see

how widespread such mechanisms are, and whether the low

variability ‘‘checkpoints’’ revealed by genetics and development

also serve to buffer probabilistic mappings.

One largely unexplored role for probabilistic behavior is, iron-

ically enough, its use to ensure deterministic outcomes in the

multicellular context. This possibility is perhaps best illustrated

by Lawton et al. (Lawton et al., 2013), in which the authors study

cell migration in the zebrafish tail bud. They found that cells

began with relatively little angular variability, but as their motion

progressed through the various zones of the embryo, the degree

of angular variability increased. Indeed, their data showed a

remarkable mixing of cells in this area. Surprisingly, their simula-

tions revealed that this disorder may in fact be critical to the reli-

able motion of cells in these zones, since very high levels of vari-

ability led to a breakdown of orderly migration, while very low

levels of disorder led to ‘‘jamming’’ in the cell migration patterns.

Thus, a moderate level of variability provided the necessary

fluidity to enable the cells to reliably move to the correct place.

A similar example involves exploiting limited variability to ensure

proper spacing of bristles on the Drosophila notum (Cohen et al.,

2010). Other similar cases may exist, but may remain hidden

from our view because of the apparent phenotypic homogeneity

masking the underlying nondeterministic behavior.

Although these studies signify the huge progress researchers

have made in finding examples of probabilistic and buffering be-

haviors, it is worth considering that thesemay represent the tip of

the iceberg. Indeed, there is a school of thought in which organ-

ismal development is poised at the edge of disorder, in which

orderly properties emerge from a gaggle of chaotic individual ac-

tors (Pelaez et al., 2015; Yuan et al., 2016). We are not sure that

the evidence necessarily supports this point of view in its en-

tirety, but we also believe it is more than just a formal possibility

and warrants further exploration.

What Are the Benefits and Detriments of ProbabilisticSingle-Cell Variability?In the previous sections, we have argued that cell-to-cell vari-

ability (due to probabilistic gene expression programs) is

prevalent and that such variability can be either propagated or

buffered to higher levels of biological organization. However, it

remains a topic of heated debate whether these phenotypic

manifestations are simply a consequence of inherently noisy

cellular processes or whether evolution has harnessed pro-

babilistic mappings as design strategies. Phrased differently,

the question is: are probabilistic maps intrinsically useful?

Conversely, what sorts of buffering maps are in place to reduce

harmful variability? Considering that evolution cannot be re-

played easily, our response remains speculative, and undoubt-

edly much probabilistic behavior is just noise, but we will high-

light examples that suggest that probabilistic behavior may

indeed have functional importance in certain settings.

Molecular Cell 62, June 2, 2016 795

Molecular Cell

Review

Probability as a Mechanism for Generating Diversity

Let us begin with the ways in which probabilistic maps can be

useful for biology. When considering this question, a good start-

ing point is to remember that organisms are wedged between

two continuous sources of variability: on one hand they are

exposed to changing and often unpredictable environments,

while on the other hand mutations will unceasingly modify their

genetic material, resulting in constant changes of gene expres-

sion levels and patterns. Given these conditions, it is likely that

the ‘‘perfect’’ mapping between genotype and phenotype is a

moving target. Not surprisingly, perhaps the best-described

rationale for why probabilistic execution may be helpful is the

generation of diversity without having to explicitly encode a reg-

ulatory scheme for every possible (and often unforeseeable)

outcome. Researchers first found examples of this sort of

behavior in bacteria, with the competence circuit in B. subtilis

(Maamar et al., 2007; Suel et al., 2006, 2007) and the Lac circuit

in E.coli (Choi et al., 2008) being the most well studied. In these

cases, the goal is to optimize fitness by committing a subpopu-

lation of cells to a particular cell fate, but not committing all the

cells. In single-celled organisms, one of the only ways to

generate such diversity if the environment is homogeneous is

to exploit probabilistic biological execution. Often these alter-

nate cell fates can be costly to generate and reduce the fitness

of a population in a given environment, but in return maximize

the fitness when the environment fluctuates, a phenomenon

referred to as bet hedging. Examples of bet hedging include bac-

terial persistence (Balaban et al., 2004) or sporulation of

B. subtilis when nutrients are limited (Veening et al., 2008).

Another intriguing recent possibility is that probabilistic gene

expression can actually lead to increased genetic diversity

through variable DNA repair (Uphoff et al., 2016).

In multicellular organisms, a similar phenomenon of diversi-

fying phenotypes in adverse environments exists. Waddington’s

classic experiments on flies showed that stress can reveal latent

genetic variability (Waddington, 1953), and work from the Lind-

quist lab has shown that Hsp90 can serve as a capacitor for

this variability (Queitsch et al., 2002; Rutherford and Lindquist,

1998). Studies in birds (Nussey et al., 2005) and plants (Nicotra

et al., 2010) have shown that variation in phenotypic plasticity

can be selected for when the environment is unfavorable. How-

ever, these examples all either rely on or cannot exclude genetic

variation as a source of variability, and a role or rationale for

nondeterministic execution is less clear. That said, in some

instances, probabilistic behavior is a part of the normal develop-

mental process (as discussed in the previous section), so it re-

mains to be seen whether probabilistic genotype-to-phenotype

mapping in multicellular organisms can be linked directly to sur-

vival in changing environments.

One intriguing use for nondeterminismmay be the selection of

the fittest (Khare and Shaulsky, 2006). Recent work inDrosophila

has shown that in the imaginal wing disk, the determination of

which cells live or die can hinge on whether the cells expressed

higher levels ofMyc, thus gaining proliferative advantages ampli-

fied by spatial interactions (Levayer et al., 2015). We consider it

worth exploring whether such mechanisms for selecting cells

that are for whatever reason the ‘‘best’’ for the particular function

are widespread.

796 Molecular Cell 62, June 2, 2016

Variability as a Regulatory Mechanism

In addition to its potential role in tolerating different environ-

ments, probabilistic behavior may also prove to be a useful

means of regulation. As a rationale, consider the difficulties com-

plex metazoa face in specifying the large number of cellular

states in the full organism. By using probability to dictate these

states, one can obtain high levels of diversity without explicitly

encoding all possibilities (the tradeoff being a lack of precise

specification). Fascinating work on probabilistic expression

and alternative splicing of clustered protocadherin genes in

mammals (Lefebvre et al., 2012) and of the Dscam gene in

Drosophila (Wojtowicz et al., 2004) has demonstrated that by ex-

pressing different variants in different neurons, these genes

regulate the self-avoidance of cells and the specificity of neural

connections. Another canonical example involves olfactory re-

ceptor neurons: each neuron expresses only one of thousands

of olfactory receptor genes, and thus every odor generates a

unique signature of neuron firing, allowing different smells to

be distinguished (Monahan and Lomvardas, 2015). Implicit in

this reasoning, however, is the notion that there is a ‘‘cost’’ to

regulation, and that at some point, the regulatory capacity of

the cell would be overwhelmed by the large number of regulators

required for direct specification.

A different benefit fromprobabilistic mappingsmay come from

the extra knobs afforded by variability (i.e., inputs to the map-

ping), which the cell may use to regulate biological activity in

novel ways. For example, the fact that transcription or signaling

as a pulsatile process opens the possibility of regulating different

aspects of these pulses, potentially leading to counterintuitive

behavior. In one beautiful example, Cai et al. showed that fre-

quency modulation of pulsatile calcium signaling allows cells to

simultaneously regulate hundreds of genes with the same

dose-response characteristics despite variation in individual

promoter responses (Cai et al., 2008). A similar example comes

from studying oscillations in NF-kB signaling. Work fromMichael

White’s group showed that this form of signaling involves a dual-

delayed negative feedback motif, where the delay in stochastic

transcription of the feedback genes is optimized to maximize

cell-to-cell heterogeneity in the phasing of the oscillations. This

increased temporal variability could then subsequently minimize

population-level fluctuations in signaling (Ashall et al., 2009; Pas-

zek et al., 2010). Transcriptional bursts also enable independent

regulatory mechanisms to co-exist, for instance, Padovan-Mer-

har et al. showed that cells achieve homeostasis through the ac-

tivity of two independent global regulatory mechanisms that in-

fluence transcriptional bursts: cell size affects burst size, and

DNA replication affects burst frequency (Padovan-Merhar

et al., 2015). Indeed, the very existence of burstsmay enable reg-

ulatory circuits that may not exist otherwise (To and Maheshri,

2010).

The Downside of Nondeterminism

While nondeterminism may have its merits, as outlined above, it

is also clear that in many instances, probabilistic behavior can be

detrimental. Fundamentally, this question is difficult to study

because perturbations that change levels of variability typically

also have other effects thatmake it difficult to ascribe phenotypic

differences to changes in variability per se (although the emer-

gence of ‘‘noise modulators’’ is an exciting development; Dar

Molecular Cell

Review

et al., 2014). Nevertheless, there is increasing evidence suggest-

ing that variability is itself a parameter subject to evolutionary

pressures, that mechanisms exists to potentially reduce noise,

and that non-genetic variability is associated with disease.

The case for selective forces acting on nondeterministic vari-

ability began with genome-wide measurements of noise in sin-

gle-celled organisms (Bar-Even et al., 2006; Newman et al.,

2006; Taniguchi et al., 2010), which revealed that housekeeping

genes exhibit less variability on average than transcription fac-

tors and other regulatory genes. High-throughput single-cell

imaging (Sigal et al., 2006) and sequencing approaches (Pado-

van-Merhar et al., 2015) have essentially corroborated the

finding that housekeeping genes are also less variable in meta-

zoan cells, although these conclusions are mostly suggestive.

There is also ample evidence to suggest that gene regulatory

networks are constructed in a manner that minimizes variability.

For example, the Gregor lab has nicely shown that in Drosophila

embryonic development, despite probabilistic transcription, the

actual transcript abundances display variability very close to

theoretical limits (Gregor et al., 2007; Little et al., 2013), and there

is some evidence that the topology of networks in bacteria has

features that reduce variability (Kollmann et al., 2005). Further-

more, perturbing normal gene regulatory networks can reveal

latent probabilistic behavior (Burga et al., 2011; Ji et al., 2013;

Raj et al., 2010; Topalidou et al., 2011), indicating that (in these

cases) precision and determinism are not the default mode of

operation, but emerge through the presence of some buffering

mechanism. Apart from the involvement of gene regulatory net-

works, such buffering can be achieved through the presence of

redundant gene regulatory elements (such as shadow en-

hancers) (Boettiger and Levine, 2009) or of redundant transcrip-

tion factors (Pioli and Weis, 2014; Stolt et al., 2004). Other

buffering mechanisms may be more complex: for example,

mice with reduced levels of the gene Trim28 show a bistable

lean-obese phenotype and, accordingly, also increased vari-

ability in behavior and gene expression. These phenotypic differ-

ences are not due to genetic or environmental causes, but rather

seem to the caused by the perturbation of an imprinted gene

network. When and how this bistable switch is triggered, how-

ever, remains enigmatic (Dalgaard et al., 2016).

Collectively, these studies suggest that probabilistic execution

of genetic programs may be deleterious for organisms and is

therefore typically suppressed. Taking an even more direct

approach to study this question, two recent papers explored

how DNA variants and selection affect gene expression vari-

ability, with the two clever studies reaching different conclusions

(Metzger et al., 2015; Wolf et al., 2015). Metzger et al. used a

normal promoter altered by both randommutations and naturally

occurring variants and measured the mean and variance of the

expression of the resulting yeast strains. They found, interest-

ingly, that while the distribution of mean expression levels was

roughly the same between the random mutants and the natural

variants, the distribution of ‘‘noise’’ took on larger values in the

randommutants than in the natural variants. InWolf et al., the au-

thors took a somewhat different approach by evolving pro-

moters, selecting only for a particular mean level of expression.

Surprisingly, they found that the noise levels of these evolved

promoters was actually lower than that of the natural promoters

and theorize that the high levels of noise facilitate regulatory evo-

lution. It is tempting, based on these results, to infer that vari-

ability is selected either for or against by evolution (depending

on the noise requirement of a given gene), although it is worth

mentioning that even if a mutation affects transcriptional vari-

ability, it may also have other effects that are the subject of selec-

tion (such as influencing mean expression levels; Ansel et al.,

2008).

Ultimately, the difficulties in specifically manipulating cellular

variability make it difficult to establish phenotypic consequences

of probabilistic behavior per se, and this remains a major chal-

lenge in the field.

Probability and Disease

In the previous examples, we painted a general picture of the

benefits and detriments of probabilistic behavior. However, un-

derstanding probabilistic phenotypes is particularly pertinent to

human health, especially considering the current emphasis on

personalized genome sequencing and precision medicine. It is

likely that at least some cases of incomplete penetrance and

variable expressivity in disease will eventually be traced to

nondeterministic mutational outcomes, thus motivating the

development of ‘‘cellular precision’’ as a complement to genetic

precision. One example is the probabilistic perturbation of

normal imprinting patterns that can lead to disease: in a recent

study, using new, single-cell, allele-specific expression mea-

surement tools in combination with epigenetic analysis, Ginart

et al. (Ginart et al., 2016) revealed that mutations to methylation

control regions can lead to probabilistic but heritable imprinting

behavior in single cells. Furthermore, this study demonstrated

that manipulating methylation levels could alter the degree of

imprinting heterogeneity, providing hints at the regulatory under-

pinnings of probabilistic behavior and how it may emerge in

abnormal contexts, such as imprinting disorders (Kalish et al.,

2014) or during in vitro fertilization (de Waal et al., 2014).

Probabilistic phenotypes can also emerge in other disease

settings, as in the case of resistance to therapies. The best-

known examples are persister bacteria (Balaban et al., 2004)

and, more recently, antibiotic resistance (El Meouche et al.,

2016). Resistance of a few cells to therapeutic agents is also

an enormous clinical issue in the treatment of cancer. Generally,

this resistance is thought to have a genetic basis, but a mounting

body of evidence suggests non-genetic mechanisms may be at

play as well (Brock et al., 2009; Gupta et al., 2011; Pisco et al.,

2013; Pisco and Huang, 2015; Sharma et al., 2010; Spencer

et al., 2009), which may inform treatment strategies (Liao et al.,

2012a, 2012b).

A Few Concluding RemarksSince we last surveyed the field (Raj and van Oudenaarden,

2008), it’s clear that the study of single-cell biology has trans-

formed inmanyways. The tools are exponentially more powerful,

bringing with them new understanding of everything from single-

molecule transcription factor binding events to evolutionary pro-

cesses guiding variability. Nevertheless, we believe that consid-

erable challenges lie waiting in the years ahead as we must now

confront some of the same questions that have remained unan-

swered through the years: what are the important molecular

events we should consider (and ignore) when studying cellular

Molecular Cell 62, June 2, 2016 797

Molecular Cell

Review

nondeterminism? How pervasive is variability through the range

of biological functions, and where does it help and where does it

hinder? What are the most informative levels of abstraction?

Exciting new research has laid tantalizing clues as to how we

might approach these problems, and we expect the next 10

years to be very interesting indeed.

ACKNOWLEDGMENTS

We thank Eduardo Torre for some early discussions in formulating this reviewand Caroline Bartman for comments and suggestions. O.S. acknowledgessupport from EMBO fellowship ALTF 691-2014. A.R. acknowledges supportfrom an NSF CAREER Award 1350601, NIH New Innovator 1DP2OD008514,NIH/NIBIB R33 EB019767, and NIH 4DN U01 HL129998. A.R. receives roy-alties from LGC/Biosearch Technologies for Stellaris RNA FISH probes. O.S.is co-founder and co-owner of English and Scientific Consulting Kft.

REFERENCES

Abranches, E., Guedes, A.M.V., Moravec, M., Maamar, H., Svoboda, P., Raj,A., and Henrique, D. (2014). Stochastic NANOG fluctuations allow mouse em-bryonic stem cells to explore pluripotency. Development 141, 2770–2779.

Achim, K., Pettit, J.-B., Saraiva, L.R., Gavriouchkina, D., Larsson, T., Arendt,D., and Marioni, J.C. (2015). High-throughput spatial mapping of single-cellRNA-seq data to tissue of origin. Nat. Biotechnol. 33, 503–509.

Albayrak, C., Jordi, C.A., Zechner, C., Lin, J., Bichsel, C.A., Khammash, M.,and Tay, S. (2016). Digital quantification of proteins and mRNA in singlemammalian cells. Mol. Cell 61, 914–924.

Amano, T., Sagai, T., Tanabe, H., Mizushina, Y., Nakazawa, H., and Shiroishi,T. (2009). Chromosomal dynamics at the Shh locus: limb bud-specific differen-tial regulation of competence and active transcription. Dev. Cell 16, 47–57.

Ansel, J., Bottin, H., Rodriguez-Beltran, C., Damon, C., Nagarajan, M., Fehr-mann, S., Francois, J., and Yvert, G. (2008). Cell-to-cell stochastic variationin gene expression is a complex genetic trait. PLoS Genet. 4, e1000049.

Arkin, A., Ross, J., and McAdams, H.H. (1998). Stochastic kinetic analysis ofdevelopmental pathway bifurcation in phage lambda-infected Escherichiacoli cells. Genetics 149, 1633–1648.

Arnaud, O., Meyer, S., Vallin, E., Beslon, G., andGandrillon, O. (2015). Temper-ature-induced variation in gene expression burst size in metazoan cells. BMCMol. Biol. 16, 20.

Ashall, L., Horton, C.A., Nelson, D.E., Paszek, P., Harper, C.V., Sillitoe, K.,Ryan, S., Spiller, D.G., Unitt, J.F., Broomhead, D.S., et al. (2009). Pulsatilestimulation determines timing and specificity of NF-kappaB-dependent tran-scription. Science 324, 242–246.

Bahar Halpern, K., and Itzkovitz, S. (2016). Single molecule approaches forquantifying transcription and degradation rates in intact mammalian tissues.Methods 98, 134–142, http://dx.doi.org/10.1016/j.ymeth.2015.11.015.

Bahar Halpern, K., Caspi, I., Lemze, D., Levy,M., Landen, S., Elinav, E., Ulitsky,I., and Itzkovitz, S. (2015a). Nuclear retention of mRNA in mammalian tissues.Cell Rep. 13, 2653–2662.

Bahar Halpern, K., Tanami, S., Landen, S., Chapal, M., Szlak, L., Hutzler, A.,Nizhberg, A., and Itzkovitz, S. (2015b). Bursty gene expression in the intactmammalian liver. Mol. Cell 58, 147–156.

Bai, L., Charvin, G., Siggia, E.D., and Cross, F.R. (2010). Nucleosome-depleted regions in cell-cycle-regulated promoters ensure reliable geneexpression in every cell cycle. Dev. Cell 18, 544–555.

Balaban, N.Q., Merrin, J., Chait, R., Kowalik, L., and Leibler, S. (2004). Bacte-rial persistence as a phenotypic switch. Science 305, 1622–1625.

Balazsi, G., van Oudenaarden, A., and Collins, J.J. (2011). Cellular decisionmaking and biological noise: from microbes to mammals. Cell 144, 910–925.

Bar-Even, A., Paulsson, J., Maheshri, N., Carmi, M., O’Shea, E., Pilpel, Y., andBarkai, N. (2006). Noise in protein expression scales with natural protein abun-dance. Nat. Genet. 38, 636–643.

798 Molecular Cell 62, June 2, 2016

Bartman, C.R., Hsu, S.C., Hsiung, C.C.-S., Raj, A., and Blobel, G.A. (2016).Enhancer regulation of transcriptional bursting parameters revealed by forcedchromatin looping. Mol. Cell 62, 237–247, http://dx.doi.org/10.1016/j.molcel.2016.03.007.

Batenchuk, C., St-Pierre, S., Tepliakova, L., Adiga, S., Szuto, A., Kabbani, N.,Bell, J.C., Baetz, K., and Kærn, M. (2011). Chromosomal position effects arelinked to sir2-mediated variation in transcriptional burst size. Biophys. J.100, L56–L58.

Battich, N., Stoeger, T., and Pelkmans, L. (2013). Image-based transcriptom-ics in thousands of single human cells at single-molecule resolution. Nat.Methods 10, 1127–1133.

Battich, N., Stoeger, T., and Pelkmans, L. (2015). Control of transcript vari-ability in single mammalian cells. Cell 163, 1596–1610.

Becskei, A., Kaufmann, B.B., and van Oudenaarden, A. (2005). Contributionsof low molecule number and chromosomal positioning to stochastic geneexpression. Nat. Genet. 37, 937–944.

Bendall, S.C., Nolan, G.P., Roederer, M., and Chattopadhyay, P.K. (2012). Adeep profiler’s guide to cytometry. Trends Immunol. 33, 323–332.

Bendall, S.C., Davis, K.L., Amir, A.D., Tadmor, M.D., Simonds, E.F., Chen, T.J.,Shenfeld, D.K., Nolan, G.P., and Pe’er, D. (2014). Single-cell trajectory detec-tion uncovers progression and regulatory coordination in human B cell devel-opment. Cell 157, 714–725.

Benzer, S. (1953). Induced synthesis of enzymes in bacteria analyzed at thecellular level. Biochim. Biophys. Acta 11, 383–395.

Bertrand, E., Chartrand, P., Schaefer, M., Shenoy, S.M., Singer, R.H., andLong, R.M. (1998). Localization of ASH1 mRNA particles in living yeast. Mol.Cell 2, 437–445.

Biase, F.H., Cao, X., and Zhong, S. (2014). Cell fate inclination within 2-cell and4-cell mouse embryos revealed by single-cell RNA sequencing. Genome Res.24, 1787–1796.

Bjornson, Z.B., Nolan, G.P., and Fantl, W.J. (2013). Single-cell mass cytometryfor analysis of immune system functional states. Curr. Opin. Immunol. 25,484–494.

Boettiger, A.N., and Levine, M. (2009). Synchronous and stochastic patterns ofgene activation in the Drosophila embryo. Science 325, 471–473.

Brock, A., Chang, H., andHuang, S. (2009). Non-genetic heterogeneity—amu-tation-independent driving force for the somatic evolution of tumours. Nat.Rev. Genet. 10, 336–342.

Brown, C.R., Mao, C., Falkovskaia, E., Jurica, M.S., and Boeger, H. (2013).Linking stochastic fluctuations in chromatin structure and gene expression.PLoS Biol. 11, e1001621.

Buenrostro, J.D., Wu, B., Litzenburger, U.M., Ruff, D., Gonzales, M.L., Snyder,M.P., Chang, H.Y., and Greenleaf, W.J. (2015). Single-cell chromatin accessi-bility reveals principles of regulatory variation. Nature 523, 486–490, http://dx.doi.org/10.1038/nature14590.

Buettner, F., Natarajan, K.N., Casale, F.P., Proserpio, V., Scialdone, A., Theis,F.J., Teichmann, S.A., Marioni, J.C., and Stegle, O. (2015). Computationalanalysis of cell-to-cell heterogeneity in single-cell RNA-sequencing data re-veals hidden subpopulations of cells. Nat. Biotechnol. 33, 155–160.

Burga, A., Casanueva, M.O., and Lehner, B. (2011). Predicting mutationoutcome from early stochastic variation in genetic interaction partners. Nature480, 250–253.

Cai, L., Friedman, N., and Xie, X.S. (2006). Stochastic protein expression in in-dividual cells at the single molecule level. Nature 440, 358–362.

Cai, L., Dalal, C.K., and Elowitz, M.B. (2008). Frequency-modulated nuclearlocalization bursts coordinate gene regulation. Nature 455, 485–490.

Chang, H.H., Hemberg, M., Barahona, M., Ingber, D.E., and Huang, S. (2008).Transcriptome-wide noise controls lineage choice in mammalian progenitorcells. Nature 453, 544–547.

Chelly, J., Concordet, J.P., Kaplan, J.C., and Kahn, A. (1989). Illegitimate tran-scription: transcription of any gene in any cell type. Proc. Natl. Acad. Sci. USA86, 2617–2621.

Molecular Cell

Review

Chen, K.H., Boettiger, A.N., Moffitt, J.R.,Wang, S., and Zhuang, X. (2015). RNAimaging. Spatially resolved, highly multiplexed RNA profiling in single cells.Science 348, aaa6090.

Choi, P.J., Cai, L., Frieda, K., and Xie, X.S. (2008). A stochastic single-moleculeevent triggers phenotype switching of a bacterial cell. Science 322, 442–446.

Chong, S., Chen, C., Ge, H., and Xie, X.S. (2014). Mechanism of transcriptionalbursting in bacteria. Cell 158, 314–326.

Chubb, J.R., Trcek, T., Shenoy, S.M., and Singer, R.H. (2006). Transcriptionalpulsing of a developmental gene. Curr. Biol. 16, 1018–1025.

Cohen, M., Georgiou, M., Stevenson, N.L., Miodownik, M., and Baum, B.(2010). Dynamic filopodia transmit intermittent Delta-Notch signaling to drivepattern refinement during lateral inhibition. Dev. Cell 19, 78–89.

Cote, A.J., McLeod, C.M., Farrell, M.J., McClanahan, P.D., Dunagin, M.C., Raj,A., and Mauck, R.L. (2016). Single-cell differences in matrix gene expressiondo not predict matrix deposition. Nat. Commun. 7, 10865, http://dx.doi.org/10.1038/ncomms10865.

Coulon, A., Ferguson, M.L., de Turris, V., Palangat, M., Chow, C.C., and Lar-son, D.R. (2014). Kinetic competition during the transcription cycle results instochastic RNA processing. eLife 3, http://dx.doi.org/10.7554/eLife.03939.

Crosetto, N., Bienko, M., and van Oudenaarden, A. (2015). Spatially resolvedtranscriptomics and beyond. Nat. Rev. Genet. 16, 57–66.

Cusanovich, D.A., Daza, R., Adey, A., Pliner, H.A., Christiansen, L., Gunder-son, K.L., Steemers, F.J., Trapnell, C., and Shendure, J. (2015). Epigenetics.Multiplex single-cell profiling of chromatin accessibility by combinatorialcellular indexing. Science 348, 910–914.

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-encodedstrategies for increasing expression with opposing effects on promoter dy-namics and transcriptional noise. Genome Res. 23, 966–976.

Dalgaard, K., Landgraf, K., Heyne, S., Lempradl, A., Longinotto, J., Gossens,K., Ruf, M., Orthofer, M., Strogantsev, R., Selvaraj, M., et al. (2016). Trim28haploinsufficiency triggers bi-stable epigenetic obesity. Cell 164, 353–364.

Dar, R.D., Razooky, B.S., Singh, A., Trimeloni, T.V., McCollum, J.M., Cox,C.D., Simpson, M.L., and Weinberger, L.S. (2012). Transcriptional burst fre-quency and burst size are equally modulated across the human genome.Proc. Natl. Acad. Sci. USA 109, 17454–17459.

Dar, R.D., Hosmane, N.N., Arkin, M.R., Siliciano, R.F., and Weinberger, L.S.(2014). Screening for noise in gene expression identifies drug synergies. Sci-ence 344, 1392–1396.

de Waal, E., Mak, W., Calhoun, S., Stein, P., Ord, T., Krapp, C., Coutifaris, C.,Schultz, R.M., and Bartolomei, M.S. (2014). In vitro culture increases the fre-quency of stochastic epigenetic errors at imprinted genes in placental tissuesfrom mouse concepti produced through assisted reproductive technologies.Biol. Reprod. 90, 22.

Dey, S.S., Foley, J.E., Limsirichai, P., Schaffer, D.V., and Arkin, A.P. (2015).Orthogonal control of expression mean and variance by epigenetic featuresat different genomic loci. Mol. Syst. Biol. 11, 806.

Dietrich, J.-E., and Hiiragi, T. (2007). Stochastic patterning in the mouse pre-implantation embryo. Development 134, 4219–4231.

Doupe, D.P., Alcolea, M.P., Roshan, A., Zhang, G., Klein, A.M., Simons, B.D.,and Jones, P.H. (2012). A single progenitor population switches behavior tomaintain and repair esophageal epithelium. Science 337, 1091–1093.

Duffy, K.R., Wellard, C.J., Markham, J.F., Zhou, J.H.S., Holmberg, R., Haw-kins, E.D., Hasbold, J., Dowling, M.R., and Hodgkin, P.D. (2012). Activation-induced B cell fates are selected by intracellular stochastic competition. Sci-ence 335, 338–341.

El Meouche, I., Siu, Y., and Dunlop, M.J. (2016). Stochastic expression of amultiple antibiotic resistance activator confers transient resistance in singlecells. Sci. Rep. 6, 19538.

Eldar, A., and Elowitz,M.B. (2010). Functional roles for noise in genetic circuits.Nature 467, 167–173.

Elf, J., Li, G.-W., and Xie, X.S. (2007). Probing transcription factor dynamics atthe single-molecule level in a living cell. Science 316, 1191–1194.

Elgin, S.C., and Reuter, G. (2013). Position-effect variegation, heterochromatinformation, and gene silencing in Drosophila. Cold Spring Harb. Perspect. Biol.5, a017780.

Elowitz, M.B., Levine, A.J., Siggia, E.D., and Swain, P.S. (2002). Stochasticgene expression in a single cell. Science 297, 1183–1186.

Felix, M.-A., and Barkoulas, M. (2015). Pervasive robustness in biological sys-tems. Nat. Rev. Genet. 16, 483–496.

Femino, A.M., Fay, F.S., Fogarty, K., and Singer, R.H. (1998). Visualization ofsingle RNA transcripts in situ. Science 280, 585–590.

Ginart, P., Kalish, J.M., Jiang, C.L., Yu, A.C., Bartolomei, M.S., and Raj, A.(2016). Visualizing allele-specific expression in single cells reveals epigeneticmosaicism in an H19 loss-of-imprinting mutant. Genes Dev. 30, 567–578.

Golding, I., Paulsson, J., Zawilski, S.M., and Cox, E.C. (2005). Real-time ki-netics of gene activity in individual bacteria. Cell 123, 1025–1036.

Goolam,M., Scialdone, A., Graham, S.J.L., Macaulay, I.C., Jedrusik, A., Hupa-lowska, A., Voet, T., Marioni, J.C., and Zernicka-Goetz,M. (2016). Heterogene-ity in Oct4 and Sox2 targets biases cell fate in 4-cell mouse embryos. Cell 165,61–74.

Gregor, T., Tank, D.W., Wieschaus, E.F., and Bialek, W. (2007). Probing thelimits to positional information. Cell 130, 153–164.

Grun, D., Kester, L., and van Oudenaarden, A. (2014). Validation of noisemodels for single-cell transcriptomics. Nat. Methods 11, 637–640.

Grunwald, D., and Singer, R.H. (2010). In vivo imaging of labelled endogenousb-actin mRNA during nucleocytoplasmic transport. Nature 467, 604–607.

Guantes, R., Rastrojo, A., Neves, R., Lima, A., Aguado, B., and Iborra, F.J.(2015). Global variability in gene expression and alternative splicing is modu-lated by mitochondrial content. Genome Res. 25, 633–644.

Guo, G., Pinello, L., Han, X., Lai, S., Shen, L., Lin, T.-W., Zou, K., Yuan, G.-C.,and Orkin, S.H. (2016). Serum-based culture conditions provoke gene expres-sion variability in mouse embryonic stem cells as revealed by single-cell anal-ysis. Cell Rep. 14, 956–965.

Gupta, P.B., Fillmore, C.M., Jiang, G., Shapira, S.D., Tao, K., Kuperwasser, C.,and Lander, E.S. (2011). Stochastic state transitions give rise to phenotypicequilibrium in populations of cancer cells. Cell 146, 633–644.

Hacisuleyman, E., Goff, L.A., Trapnell, C., Williams, A., Henao-Mejia, J., Sun,L., McClanahan, P., Hendrickson, D.G., Sauvageau, M., Kelley, D.R., et al.(2014). Topological organization of multichromosomal regions by the long in-tergenic noncoding RNA Firre. Nat. Struct. Mol. Biol. 21, 198–206.

Halstead, J.M., Lionnet, T., Wilbertz, J.H., Wippich, F., Ephrussi, A., Singer,R.H., and Chao, J.A. (2015). Translation. An RNA biosensor for imaging the firstround of translation from single cells to living animals. Science 347, 1367–1671.

Hammar, P., Leroy, P., Mahmutovic, A., Marklund, E.G., Berg, O.G., and Elf, J.(2012). The lac repressor displays facilitated diffusion in living cells. Science336, 1595–1598.

Hansen, C.H., and van Oudenaarden, A. (2013). Allele-specific detection ofsingle mRNA molecules in situ. Nat. Methods 10, 869–871.

Hensel, Z., Feng, H., Han, B., Hatem, C., Wang, J., and Xiao, J. (2012). Sto-chastic expression dynamics of a transcription factor revealed by single-mole-cule noise analysis. Nat. Struct. Mol. Biol. 19, 797–802.

Inaba, M., and Yamashita, Y.M. (2012). Asymmetric stem cell division: preci-sion for robustness. Cell Stem Cell 11, 461–469.

Itzkovitz, S., and van Oudenaarden, A. (2011). Validating transcripts withprobes and imaging technology. Nat. Methods 8 (4, Suppl), S12–S19.

Ji, N., Middelkoop, T.C., Mentink, R.A., Betist, M.C., Tonegawa, S., Mooijman,D., Korswagen, H.C., and van Oudenaarden, A. (2013). Feedback control ofgene expression variability in the Caenorhabditis elegans Wnt pathway. Cell155, 869–880.

Molecular Cell 62, June 2, 2016 799

Molecular Cell

Review

Jiang, Y.J., Aerne, B.L., Smithers, L., Haddon, C., Ish-Horowicz, D., and Lewis,J. (2000). Notch signalling and the synchronization of the somite segmentationclock. Nature 408, 475–479.

Johnston, I.G., Gaal, B., Neves, R.P., Enver, T., Iborra, F.J., and Jones, N.S.(2012). Mitochondrial variability as a source of extrinsic cellular noise. PLoSComput. Biol. 8, e1002416.

Jones, D.L., Brewster, R.C., and Phillips, R. (2014). Promoter architecture dic-tates cell-to-cell variability in gene expression. Science 346, 1533–1536.

Junker, J.P., Noel, E.S., Guryev, V., Peterson, K.A., Shah, G., Huisken, J.,McMahon, A.P., Berezikov, E., Bakkers, J., and van Oudenaarden, A. (2014).Genome-wide RNA tomography in the zebrafish embryo. Cell 159, 662–675.

Kalish, J.M., Jiang, C., andBartolomei,M.S. (2014). Epigenetics and imprintingin human disease. Int. J. Dev. Biol. 58, 291–298.

Kalmar, T., Lim, C., Hayward, P., Munoz-Descalzo, S., Nichols, J., Garcia-Ojalvo, J., and Martinez Arias, A. (2009). Regulated fluctuations in nanogexpression mediate cell fate decisions in embryonic stem cells. PLoS Biol. 7,e1000149.

Katz, Z.B., English, B.P., Lionnet, T., Yoon, Y.J., Monnier, N., Ovryn, B., Bathe,M., and Singer, R.H. (2016). Mapping translation ‘hot-spots’ in live cells bytracking single molecules of mRNA and ribosomes. eLife 5, e10415, http://dx.doi.org/10.7554/eLife.10415.

Ke, R., Mignardi, M., Pacureanu, A., Svedlund, J., Botling, J., Wahlby, C., andNilsson, M. (2013). In situ sequencing for RNA analysis in preserved tissue andcells. Nat. Methods 10, 857–860.

Kempe, H., Schwabe, A., Cremazy, F., Verschure, P.J., and Bruggeman, F.J.(2015). The volumes and transcript counts of single cells reveal concentrationhomeostasis and capture biological noise. Mol. Biol. Cell 26, 797–804.

Kepler, T.B., and Elston, T.C. (2001). Stochasticity in transcriptional regulation:origins, consequences, and mathematical representations. Biophys. J. 81,3116–3136.

Khare, A., and Shaulsky, G. (2006). First among equals: competition betweengenetically identical cells. Nat. Rev. Genet. 7, 577–583.

Klein, A.M., Mazutis, L., Akartuna, I., Tallapragada, N., Veres, A., Li, V., Pesh-kin, L., Weitz, D.A., and Kirschner, M.W. (2015). Droplet barcoding for single-cell transcriptomics applied to embryonic stem cells. Cell 161, 1187–1201.

Ko, M.S., Nakauchi, H., and Takahashi, N. (1990). The dose dependence ofglucocorticoid-inducible gene expression results from changes in the numberof transcriptionally active templates. EMBO J. 9, 2835–2842.

Kollmann, M., Løvdok, L., Bartholome, K., Timmer, J., and Sourjik, V. (2005).Design principles of a bacterial signalling network. Nature 438, 504–507.

Krieger, T., and Simons, B.D. (2015). Dynamic stem cell heterogeneity. Devel-opment 142, 1396–1406.

Larson, D.R., Zenklusen, D., Wu, B., Chao, J.A., and Singer, R.H. (2011). Real-time observation of transcription initiation and elongation on an endogenousyeast gene. Science 332, 475–478.

Lawton, A.K., Nandi, A., Stulberg, M.J., Dray, N., Sneddon, M.W., Pontius, W.,Emonet, T., and Holley, S.A. (2013). Regulated tissue fluidity steers zebrafishbody elongation. Development 140, 573–582.

Lee, J.H., Daugharthy, E.R., Scheiman, J., Kalhor, R., Yang, J.L., Ferrante,T.C., Terry, R., Jeanty, S.S.F., Li, C., Amamoto, R., et al. (2014). Highly multi-plexed subcellular RNA sequencing in situ. Science 343, 1360–1363.

Lefebvre, J.L., Kostadinov, D., Chen, W.V., Maniatis, T., and Sanes, J.R.(2012). Protocadherins mediate dendritic self-avoidance in the mammaliannervous system. Nature 488, 517–521.

Levayer, R., Hauert, B., and Moreno, E. (2015). Cell mixing induced by myc isrequired for competitive tissue invasion and destruction. Nature 524, 476–480.

Levesque, M.J., and Raj, A. (2013). Single-chromosome transcriptionalprofiling reveals chromosomal gene expression regulation. Nat. Methods 10,246–248.

Levesque, M.J., Ginart, P., Wei, Y., and Raj, A. (2013). Visualizing SNVs toquantify allele-specific expression in single cells. Nat. Methods 10, 865–867.

800 Molecular Cell 62, June 2, 2016

Levsky, J.M., Shenoy, S.M., Pezo, R.C., and Singer, R.H. (2002). Single-cellgene expression profiling. Science 297, 836–840.

Liao, D., Estevez-Salmeron, L., and Tlsty, T.D. (2012a). Conceptualizing a toolto optimize therapy based on dynamic heterogeneity. Phys. Biol. 9, 065005.

Liao, D., Estevez-Salmeron, L., and Tlsty, T.D. (2012b). Generalized principlesof stochasticity can be used to control dynamic heterogeneity. Phys. Biol. 9,065006.

Little, S.C., Tikhonov, M., and Gregor, T. (2013). Precise developmental geneexpression arises from globally stochastic transcriptional activity. Cell 154,789–800.

Liu, Z., Legant, W.R., Chen, B.-C., Li, L., Grimm, J.B., Lavis, L.D., Betzig, E.,and Tjian, R. (2014). 3D imaging of Sox2 enhancer clusters in embryonicstem cells. eLife 3, e04236.

Lubeck, E., and Cai, L. (2012). Single-cell systems biology by super-resolutionimaging and combinatorial labeling. Nat. Methods 9, 743–748.

Lubeck, E., Coskun, A.F., Zhiyentayev, T., Ahmad, M., and Cai, L. (2014). Sin-gle-cell in situ RNA profiling by sequential hybridization. Nat. Methods 11,360–361.

Lyon, M.F. (1961). Gene action in the X-chromosome of the mouse (Mus mus-culus L.). Nature 190, 372–373.

Maamar, H., Raj, A., and Dubnau, D. (2007). Noise in gene expression deter-mines cell fate in Bacillus subtilis. Science 317, 526–529.

Maamar, H., Cabili, M.N., Rinn, J., and Raj, A. (2013). linc-HOXA1 is a noncod-ing RNA that represses Hoxa1 transcription in cis. Genes Dev. 27, 1260–1271.

Macosko, E.Z., Basu, A., Satija, R., Nemesh, J., Shekhar, K., Goldman, M., Tir-osh, I., Bialas, A.R., Kamitaki, N., Martersteck, E.M., et al. (2015). Highlyparallel genome-wide expression profiling of individual cells using nanoliterdroplets. Cell 161, 1202–1214.

Marinov, G.K., Williams, B.A., McCue, K., Schroth, G.P., Gertz, J., Myers,R.M., and Wold, B.J. (2014). From single-cell to cell-pool transcriptomes: sto-chasticity in gene expression and RNA splicing. Genome Res. 24, 496–510.

Martinez Arias, A., Nichols, J., and Schroter, C. (2013). A molecular basis fordevelopmental plasticity in early mammalian embryos. Development 140,3499–3510.

Masamizu, Y., Ohtsuka, T., Takashima, Y., Nagahara, H., Takenaka, Y., Yosh-ikawa, K., Okamura, H., and Kageyama, R. (2006). Real-time imaging of the so-mite segmentation clock: revelation of unstable oscillators in the individualpresomitic mesoderm cells. Proc. Natl. Acad. Sci. USA 103, 1313–1318.

McAdams, H.H., and Arkin, A. (1997). Stochastic mechanisms in gene expres-sion. Proc. Natl. Acad. Sci. USA 94, 814–819.

McLaren, A., and Michie, D. (1958). Factors affecting vertebral variation inmice. 4. Experimental proof of the uterine basis of a maternal effect.J. Embryol. Exp. Morphol. 6, 645–659.

Mellis, I.A., and Raj, A. (2015). Half dozen of one, six billion of the other: whatcan small- and large-scale molecular systems biology learn from one another?Genome Res. 25, 1466–1472.

Metzger, B.P.H., Yuan, D.C., Gruber, J.D., Duveau, F., and Wittkopp, P.J.(2015). Selection on noise constrains variation in a eukaryotic promoter. Nature521, 344–347.

Misteli, T. (2007). Beyond the sequence: cellular organization of genome func-tion. Cell 128, 787–800.

Monahan, K., and Lomvardas, S. (2015). Monoallelic expression of olfactoryreceptors. Annu. Rev. Cell Dev. Biol. 31, 721–740.

Nair, G., Abranches, E., Guedes, A.M.V., Henrique, D., and Raj, A. (2015). Het-erogeneous lineagemarker expression in naive embryonic stem cells is mostlydue to spontaneous differentiation. Sci. Rep. 5, 13339.

Newman, J.R.S., Ghaemmaghami, S., Ihmels, J., Breslow, D.K., Noble, M.,DeRisi, J.L., and Weissman, J.S. (2006). Single-cell proteomic analysis of S.cerevisiae reveals the architecture of biological noise. Nature 441, 840–846.

Molecular Cell

Review

Nicotra, A.B., Atkin, O.K., Bonser, S.P., Davidson, A.M., Finnegan, E.J., Math-esius, U., Poot, P., Purugganan, M.D., Richards, C.L., Valladares, F., and vanKleunen, M. (2010). Plant phenotypic plasticity in a changing climate. TrendsPlant Sci. 15, 684–692.

Noordermeer, D., deWit, E., Klous, P., van deWerken, H., Simonis, M., Lopez-Jones, M., Eussen, B., de Klein, A., Singer, R.H., and de Laat, W. (2011). Var-iegated gene expression caused by cell-specific long-range DNA interactions.Nat. Cell Biol. 13, 944–951.

Novick, A., and Weiner, M. (1957). Enzyme induction as an all-or-none phe-nomenon. Proc. Natl. Acad. Sci. USA 43, 553–566.

Nussey, D.H., Postma, E., Gienapp, P., and Visser, M.E. (2005). Selection onheritable phenotypic plasticity in a wild bird population. Science 310, 304–306.

Octavio, L.M., Gedeon, K., and Maheshri, N. (2009). Epigenetic and conven-tional regulation is distributed among activators of FLO11 allowing tuning ofpopulation-level heterogeneity in its expression. PLoS Genet. 5, e1000673.

Ohnishi, Y., Huber, W., Tsumura, A., Kang, M., Xenopoulos, P., Kurimoto, K.,Ole�s, A.K., Arauzo-Bravo, M.J., Saitou, M., Hadjantonakis, A.-K., and Hiiragi,T. (2014). Cell-to-cell expression variability followed by signal reinforcementprogressively segregates early mouse lineages. Nat. Cell Biol. 16, 27–37.

Ozbudak, E.M., Thattai, M., Kurtser, I., Grossman, A.D., and van Oudenaar-den, A. (2002). Regulation of noise in the expression of a single gene. Nat.Genet. 31, 69–73.

Padovan-Merhar, O., Nair, G.P., Biaesch, A.G., Mayer, A., Scarfone, S., Foley,S.W.,Wu, A.R., Churchman, L.S., Singh, A., and Raj, A. (2015). Single mamma-lian cells compensate for differences in cellular volume and DNA copy numberthrough independent global transcriptional mechanisms. Mol. Cell 58,339–352.

Paszek, P., Ryan, S., Ashall, L., Sillitoe, K., Harper, C.V., Spiller, D.G., Rand,D.A., and White, M.R.H. (2010). Population robustness arising from cellularheterogeneity. Proc. Natl. Acad. Sci. USA 107, 11644–11649.

Paulsson, J. (2005). Models of stochastic gene expression. Phys. Life Rev. 2,157–175.

Peccoud, J., and Ycart, B. (1995). Markovian modelling of gene product syn-thesis. Theor. Popul. Biol. 48, 222–234.

Pelaez, N., Gavalda-Miralles, A., Wang, B., Navarro, H.T., Gudjonson, H., Re-bay, I., Dinner, A.R., Katsaggelos, A.K., Amaral, L.A., and Carthew, R.W.(2015). Dynamics and heterogeneity of a fate determinant during transition to-wards cell differentiation. eLife 4, e08924, http://dx.doi.org/10.7554/eLife.08924.

Pina, C., Fugazza, C., Tipping, A.J., Brown, J., Soneji, S., Teles, J., Peterson,C., and Enver, T. (2012). Inferring rules of lineage commitment in haematopoi-esis. Nat. Cell Biol. 14, 287–294.

Pioli, P.D., and Weis, J.H. (2014). Snail transcription factors in hematopoieticcell development: a model of functional redundancy. Exp. Hematol. 42,425–430.

Pisco, A.O., and Huang, S. (2015). Non-genetic cancer cell plasticity and ther-apy-induced stemness in tumour relapse: ‘what does not kill me strengthensme’. Br. J. Cancer 112, 1725–1732.

Pisco, A.O., Brock, A., Zhou, J., Moor, A., Mojtahedi, M., Jackson, D., andHuang, S. (2013). Non-Darwinian dynamics in therapy-induced cancer drugresistance. Nat. Commun. 4, 2467.

Plusa, B., Piliszek, A., Frankenberg, S., Artus, J., and Hadjantonakis, A.-K.(2008). Distinct sequential cell behaviours direct primitive endoderm formationin the mouse blastocyst. Development 135, 3081–3091.

Queitsch, C., Sangster, T.A., and Lindquist, S. (2002). Hsp90 as a capacitor ofphenotypic variation. Nature 417, 618–624.

Raj, A., and van Oudenaarden, A. (2008). Nature, nurture, or chance: stochas-tic gene expression and its consequences. Cell 135, 216–226.

Raj, A., and van Oudenaarden, A. (2009). Single-molecule approaches to sto-chastic gene expression. Annu. Rev. Biophys. 38, 255–270.

Raj, A., Peskin, C.S., Tranchina, D., Vargas, D.Y., and Tyagi, S. (2006). Sto-chastic mRNA synthesis in mammalian cells. PLoS Biol. 4, e309.

Raj, A., van den Bogaard, P., Rifkin, S.A., van Oudenaarden, A., and Tyagi, S.(2008). Imaging individual mRNA molecules using multiple singly labeledprobes. Nat. Methods 5, 877–879.

Raj, A., Rifkin, S.A., Andersen, E., and van Oudenaarden, A. (2010). Variabilityin gene expression underlies incomplete penetrance. Nature 463, 913–918.

Raser, J.M., and O’Shea, E.K. (2004). Control of stochasticity in eukaryoticgene expression. Science 304, 1811–1814.

Rosenfeld, N., Young, J.W., Alon, U., Swain, P.S., and Elowitz, M.B. (2005).Gene regulation at the single-cell level. Science 307, 1962–1965.

Rutherford, S.L., and Lindquist, S. (1998). Hsp90 as a capacitor formorpholog-ical evolution. Nature 396, 336–342.

Satija, R., Farrell, J.A., Gennert, D., Schier, A.F., and Regev, A. (2015). Spatialreconstruction of single-cell gene expression data. Nat. Biotechnol. 33,495–502.

Schoenfelder, S., Clay, I., and Fraser, P. (2010). The transcriptional interac-tome: gene expression in 3D. Curr. Opin. Genet. Dev. 20, 127–133.

Schrodinger, E. (1944). What is Life? (Cambridge University Press).

Senecal, A., Munsky, B., Proux, F., Ly, N., Braye, F.E., Zimmer, C., Mueller, F.,and Darzacq, X. (2014). Transcription factors modulate c-Fos transcriptionalbursts. Cell Rep. 8, 75–83.

Sepulveda, L.A., Xu, H., Zhang, J., Wang, M., and Golding, I. (2016). Measure-ment of gene regulation in individual cells reveals rapid switching between pro-moter states. Science 351, 1218–1222.

Shah, K., and Tyagi, S. (2013). Barriers to transmission of transcriptional noisein a c-fos c-jun pathway. Mol. Syst. Biol. 9, 687.

Sharma, S.V., Lee, D.Y., Li, B., Quinlan, M.P., Takahashi, F., Maheswaran, S.,McDermott, U., Azizian, N., Zou, L., Fischbach, M.A., et al. (2010). A chro-matin-mediated reversible drug-tolerant state in cancer cell subpopulations.Cell 141, 69–80.

Sigal, A., Milo, R., Cohen, A., Geva-Zatorsky, N., Klein, Y., Liron, Y., Rosenfeld,N., Danon, T., Perzov, N., and Alon, U. (2006). Variability andmemory of proteinlevels in human cells. Nature 444, 643–646.

Singer, Z.S., Yong, J., Tischler, J., Hackett, J.A., Altinok, A., Surani, M.A., Cai,L., and Elowitz, M.B. (2014). Dynamic heterogeneity and DNA methylation inembryonic stem cells. Mol. Cell 55, 319–331.

Singh, A., Razooky, B., Cox, C.D., Simpson, M.L., andWeinberger, L.S. (2010).Transcriptional bursting from the HIV-1 promoter is a significant source of sto-chastic noise in HIV-1 gene expression. Biophys. J. 98, L32–L34.

Skinner, S.O., Xu, H., Nagarkar-Jaiswal, S., Freire, P.R., Zwaka, T.P., andGolding, I. (2016). Single-cell analysis of transcription kinetics across the cellcycle. eLife 5, e12175, http://dx.doi.org/10.7554/eLife.12175.

Smith, A. (2013). Nanog heterogeneity: tilting at windmills? Cell Stem Cell 13,6–7.

Spencer, S.L., Gaudet, S., Albeck, J.G., Burke, J.M., and Sorger, P.K. (2009).Non-genetic origins of cell-to-cell variability in TRAIL-induced apoptosis. Na-ture 459, 428–432, http://dx.doi.org/10.1038/nature08012.

Splinter, E., and de Laat, W. (2011). The complex transcription regulatory land-scape of our genome: control in three dimensions. EMBO J. 30, 4345–4355.

Spudich, J.L., and Koshland, D.E., Jr. (1976). Non-genetic individuality:chance in the single cell. Nature 262, 467–471.

St-Pierre, F., and Endy, D. (2008). Determination of cell fate selection duringphage lambda infection. Proc. Natl. Acad. Sci. USA 105, 20705–20710.

Stelzer, Y., Shivalila, C.S., Soldner, F., Markoulaki, S., and Jaenisch, R. (2015).Tracing dynamic changes of DNA methylation at single-cell resolution. Cell163, 218–229.

Stolt, C.C., Lommes, P., Friedrich, R.P., and Wegner, M. (2004). Transcriptionfactors Sox8 and Sox10 perform non-equivalent roles during oligodendrocytedevelopment despite functional redundancy. Development 131, 2349–2358.

Molecular Cell 62, June 2, 2016 801

Molecular Cell

Review

Suel, G.M., Garcia-Ojalvo, J., Liberman, L.M., and Elowitz, M.B. (2006). Anexcitable gene regulatory circuit induces transient cellular differentiation. Na-ture 440, 545–550.

Suel, G.M., Kulkarni, R.P., Dworkin, J., Garcia-Ojalvo, J., and Elowitz, M.B.(2007). Tunability and noise dependence in differentiation dynamics. Science315, 1716–1719.

Suter, D.M., Molina, N., Gatfield, D., Schneider, K., Schibler, U., and Naef, F.(2011). Mammalian genes are transcribed with widely different bursting ki-netics. Science 332, 472–474.

Takahashi, K., and Yamanaka, S. (2006). Induction of pluripotent stem cellsfrom mouse embryonic and adult fibroblast cultures by defined factors. Cell126, 663–676.

Taniguchi, Y., Choi, P.J., Li, G.-W., Chen, H., Babu, M., Hearn, J., Emili, A., andXie, X.S. (2010). Quantifying E. coli proteome and transcriptome with single-molecule sensitivity in single cells. Science 329, 533–538.

Tay, S., Hughey, J.J., Lee, T.K., Lipniacki, T., Quake, S.R., and Covert, M.W.(2010). Single-cell NF-kappaB dynamics reveal digital activation and analogueinformation processing. Nature 466, 267–271.

Thattai, M., and van Oudenaarden, A. (2001). Intrinsic noise in gene regulatorynetworks. Proc. Natl. Acad. Sci. USA 98, 8614–8619.

To, T.-L., andMaheshri, N. (2010). Noise can induce bimodality in positive tran-scriptional feedback loops without bistability. Science 327, 1142–1145.

Topalidou, I., van Oudenaarden, A., and Chalfie, M. (2011). Caenorhabditis el-egans aristaless/Arx gene alr-1 restricts variable gene expression. Proc. Natl.Acad. Sci. USA 108, 4063–4068.

Trapnell, C. (2015). Defining cell types and states with single-cell genomics.Genome Res. 25, 1491–1498.

Trapnell, C., Cacchiarelli, D., Grimsby, J., Pokharel, P., Li, S., Morse, M., Len-non, N.J., Livak, K.J., Mikkelsen, T.S., and Rinn, J.L. (2014). The dynamics andregulators of cell fate decisions are revealed by pseudotemporal ordering ofsingle cells. Nat. Biotechnol. 32, 381–386.

Tsiairis, C.D., and Aulehla, A. (2016). Self-organization of embryonic geneticoscillators into spatiotemporal wave patterns. Cell 164, 656–667.

Umulis, D., O’Connor, M.B., and Othmer, H.G. (2008). Robustness of embry-onic spatial patterning in Drosophila melanogaster. In Current Topics in Devel-opmental Biology, S. Schnell, P.K. Maini, S.A. Newman, and T.J. Newman,eds. (Academic Press), pp. 65–111.

Uphoff, S., Lord, N.D., Okumus, B., Potvin-Trottier, L., Sherratt, D.J., andPaulsson, J. (2016). Stochastic activation of a DNA damage response causescell-to-cell mutation rate variation. Science 351, 1094–1097.

Veening, J.-W., Stewart, E.J., Berngruber, T.W., Taddei, F., Kuipers, O.P., andHamoen, L.W. (2008). Bet-hedging and epigenetic inheritance in bacterial celldevelopment. Proc. Natl. Acad. Sci. USA 105, 4393–4398.

Vinuelas, J., Kaneko, G., Coulon, A., Vallin, E., Morin, V., Mejia-Pous, C., Ku-piec, J.-J., Beslon, G., and Gandrillon, O. (2013). Quantifying the contributionof chromatin dynamics to stochastic gene expression reveals long, locus-dependent periods between transcriptional bursts. BMC Biol. 11, 15.

Vogt, G. (2015). Stochastic developmental variation, an epigenetic source ofphenotypic diversity with far-reaching biological consequences. J. Biosci.40, 159–204.

802 Molecular Cell 62, June 2, 2016

Waddington, C.H. (1953). Genetic assimilation of an acquired character. Evo-lution 7, 118–126.

Waks, Z., Klein, A.M., and Silver, P.A. (2011). Cell-to-cell variability of alterna-tive RNA splicing. Mol. Syst. Biol. 7, 506.

Weinberger, L., Voichek, Y., Tirosh, I., Hornung, G., Amit, I., and Barkai, N.(2012). Expression noise and acetylation profiles distinguish HDAC functions.Mol. Cell 47, 193–202.

Wernet, M.F., Mazzoni, E.O., Celik, A., Duncan, D.M., Duncan, I., and Desplan,C. (2006). Stochastic spineless expression creates the retinal mosaic forcolour vision. Nature 440, 174–180.

White, M.D., Angiolini, J.F., Alvarez, Y.D., Kaur, G., Zhao, Z.W., Mocskos, E.,Bruno, L., Bissiere, S., Levi, V., and Plachta, N. (2016). Long-lived binding ofSox2 to DNA predicts cell fate in the four-cell mouse embryo. Cell 165, 75–87.

Wojtowicz, W.M., Flanagan, J.J., Millard, S.S., Zipursky, S.L., and Clemens,J.C. (2004). Alternative splicing of Drosophila Dscam generates axon guidancereceptors that exhibit isoform-specific homophilic binding. Cell 118, 619–633.

Wolf, L., Silander, O.K., and van Nimwegen, E. (2015). Expression noise facil-itates the evolution of gene regulation. eLife 4, http://dx.doi.org/10.7554/eLife.05856.

Wu, A.R., Neff, N.F., Kalisky, T., Dalerba, P., Treutlein, B., Rothenberg, M.E.,Mburu, F.M., Mantalas, G.L., Sim, S., Clarke, M.F., and Quake, S.R. (2014).Quantitative assessment of single-cell RNA-sequencing methods. Nat.Methods 11, 41–46.

Wu, B., Buxbaum, A.R., Katz, Z.B., Yoon, Y.J., and Singer, R.H. (2015). Quan-tifying protein-mRNA interactions in single live cells. Cell 162, 211–220.

Xu, H., Sepulveda, L.A., Figard, L., Sokac, A.M., and Golding, I. (2015).Combining protein and mRNA quantification to decipher transcriptional regu-lation. Nat. Methods 12, 739–742.

Yu, J., Xiao, J., Ren, X., Lao, K., and Xie, X.S. (2006). Probing gene expressionin live cells, one protein molecule at a time. Science 311, 1600–1603.

Yuan, L., Chan, G.C., Beeler, D., Janes, L., Spokes, K.C., Dharaneeswaran, H.,Mojiri, A., Adams, W.J., Sciuto, T., Garcia-Cardena, G., et al. (2016). A role ofstochastic phenotype switching in generating mosaic endothelial cell hetero-geneity. Nat. Commun. 7, 10160.

Yunger, S., Rosenfeld, L., Garini, Y., and Shav-Tal, Y. (2010). Single-allele anal-ysis of transcription kinetics in living mammalian cells. Nat. Methods 7,631–633.

Zeng, L., Skinner, S.O., Zong, C., Sippy, J., Feiss, M., and Golding, I. (2010).Decision making at a subcellular level determines the outcome of bacterio-phage infection. Cell 141, 682–691.

Zenklusen, D., Larson, D.R., and Singer, R.H. (2008). Single-RNA counting re-veals alternative modes of gene expression in yeast. Nat. Struct. Mol. Biol. 15,1263–1271.

Zenobi, R. (2013). Single-cell metabolomics: analytical and biological perspec-tives. Science 342, 1243259.

Zopf, C.J., Quinn, K., Zeidman, J., and Maheshri, N. (2013). Cell-cycle depen-dence of transcription dominates noise in gene expression. PLoS Comput.Biol. 9, e1003161.

Zwijnenburg, P.J.G., Meijers-Heijboer, H., and Boomsma, D.I. (2010). Identicalbut not the same: the value of discordant monozygotic twins in geneticresearch. Am. J. Med. Genet. B. Neuropsychiatr. Genet. 153B, 1134–1149.