Yeast commits approximately 76% of its energy budget to protein synthesis and the efficiency and control of this process are accordingly critical to organism growth and fitness. We now have detailed genetic, biochemical and biophysical knowledge of the components of the eukaryotic translation machinery. However, these kinds of information do not, in themselves, give us a satisfactory picture of how the overall system is controlled. This is where quantitative system analysis can enable a step-change in our understanding of biological resource management and how this relates to cell physiology and evolution. An important aspect of this more system-oriented approach to translational control is the inherent heterogeneity of cell populations that is generated by gene expression noise. In this short review, we address the fact that, although the vast majority of our knowledge of the translation machinery is based on experimental analysis of samples that each contain hundreds of millions of cells, in reality every cell is unique in terms of its composition and control properties. We have entered a new era in which research into the heterogeneity of cell systems promises to provide answers to many (previously unanswerable) questions about cell physiology and evolution.

Introduction

Living cells have to coordinate the activities and interactions of thousands of molecular species. Elucidation of the rate control properties and mechanisms underpinning biological functionality represents one of the great challenges of modern biology and will require the sustained application of advanced experimental and computational strategies. In this short review, we consider progress towards understanding gene expression, with a particular focus on the mRNA translation machinery. A very interesting dimension to this challenge is the relationship between the average behaviour of large numbers of cells, as studied in traditional biochemistry, and the specific properties of cellular machineries within an individual cell. Work in this area promises to yield fascinating insight into the principles of life at small molecular numbers and how these relate to cell physiology and evolution.

Rate control analysis

For many decades, the classical biochemical approach to analysing metabolic pathways was primarily via the characterization of isolated enzymes. One of the key concepts of rate control in these pathways was that there should exist a rate-limiting step that dictates the overall pathway flux. However, we now know that, since the reaction steps on a pathway are tightly interconnected, overall pathway flux is a function of multiple steps. A more appropriate approach is that of metabolic control analysis, in which it is assumed that control is shared between the component steps and each step's contribution to pathway flux control can be quantified and expressed as a flux control coefficient [1]. The flux control coefficient of an enzyme is a system property that cannot be related to the enzyme in isolation, although there are links between the enzyme kinetic properties and its potential for flux control. The quantitative measure that relates the flux coefficient to the influence of effectors external to the pathway is called the response coefficient [2,3]. It is frequently the case that precise kinetic data are not available for all of the individual pathway enzymes. However, response coefficients can still be measured in vivo for the corresponding steps of a pathway (see later). Overall, data on the kinetics and control characteristics of a pathway are often incomplete, but computational models can still enable researchers to gain significant insight into rate control in such situations [4].

The idea of representing the translation pathway in mathematical terms is not new. The very first model of this process was constructed almost five decades ago [5] (for more details see a review on computational models of translation in the study by von der Haar [6]). The main obstacle was to obtain experimental data to complement computation, so those models were, in most cases, purely theoretical. The development of the ribosome profiling technique [7] provided valuable data sets upon which quite a few new models were based. Those models demonstrated that the translation rate depends on availability of mRNA and free ribosomes, the codon composition of an mRNA molecule and the concentrations and availability of tRNA species [810].

Recent work has seen the application of the approach of metabolic control analysis to the translation pathway through the construction of a highly parameterized model of Saccharomyces cerevisiae translation [11]. An experimental method was developed to determine response coefficients, using the simplified definition of the relationship between a given change in concentration (activity) of a given factor and the resulting change in overall translation rate. This study showed that control in the translation pathway in exponentially growing yeast cells is not limited to the initiation step and that it is shared between multiple factors from all three stages of eukaryotic mRNA translation: initiation, elongation and termination (Figure 1A). The biggest influence over flux, indicated by the largest response coefficients, is exercised by the elongation step. The experimental data were incorporated into a model of translation constructed using ordinary differential equations. The resulting model was able to generate good fits to the experimental data (Figure 1B). It also highlighted the dependence of translation on the availability of mRNA and of free ribosomes; changes in the amount of free ribosomes over a range of concentrations close to the one expected in living cells were predicted to exert a strong influence on translation rate (Figure 1C and 1D). If the elongation step was slowed down in the model, it predicted the accumulation of ribosomes on mRNA molecules, which in turn depleted the free ribosome pool available for translation initiation. Overall, the 10 translation factors that were observed to have high response coefficients did not fit any predictable pattern and their influence over control would be hard to deduce in any other way. This is therefore a good example of how a sophisticated mathematical model can provide new, unintuitive insights into the control of a highly complex system [11].

Rate control in the yeast translation machinery

Figure 1
Rate control in the yeast translation machinery

Summary of main conclusions from a comprehensive rate control study of yeast translation [11]: (A). Bar chart showing the non-zero RJ1 values (colour scheme: yellow, initiation; orange, elongation; green, termination factors). Predictions of the in silico translation model: (B). Comparison of experimental global protein synthesis rate data (green) and the equivalent rate data calculated by a digital translation pathway model (blue), in this case predicting the relationship between total flux through the pathway and the intracellular abundance of eukaryotic initiation factor 4A (eIF4A). The model also predicts the dependence of protein synthesis rate on intracellular mRNA abundance (C) and on ribosome content per cell (D).

Figure 1
Rate control in the yeast translation machinery

Summary of main conclusions from a comprehensive rate control study of yeast translation [11]: (A). Bar chart showing the non-zero RJ1 values (colour scheme: yellow, initiation; orange, elongation; green, termination factors). Predictions of the in silico translation model: (B). Comparison of experimental global protein synthesis rate data (green) and the equivalent rate data calculated by a digital translation pathway model (blue), in this case predicting the relationship between total flux through the pathway and the intracellular abundance of eukaryotic initiation factor 4A (eIF4A). The model also predicts the dependence of protein synthesis rate on intracellular mRNA abundance (C) and on ribosome content per cell (D).

Gene expression stochasticity introduces an additional level of complexity

Until recently, biochemical studies of translation have been performed almost exclusively on large cell populations, generating bulk data obtained using hundreds of millions of cells. However, biological processes such as translation occur within individual cells and thus the averaging that is imposed by analysing bulk samples obscures cell-to-cell variations. As a result, it has only recently become apparent that there are previously unsuspected levels of heterogeneity in isogenic cell populations. This step-change in the appreciation of the importance of cell-to-cell heterogeneity has been driven by the rapid development of single-cell analytical techniques [12,13].

Stochasticity in cells arises from the fact that biochemical pathways often involve a small number of molecules undergoing Brownian motion that need to interact together to participate in a specific reaction. Such interactions are the results of random encounters between those molecules, giving rise to random fluctuations. Gene expression is central to all biochemical processes and particularly susceptible to stochasticity as it typically starts from the binding of transcription factors (more or less abundant) that allow transcription to initiate at a low copy number of specific gene promoters (one or two per cell). In addition, all subsequent steps in gene expression, leading to synthesis and degradation of mRNA and protein molecules, are also subject to random variations. This gene expression ‘noise’ results in mRNA and protein number fluctuations both as a function of time within a single cell and at any given time point from cell to cell within a genetically identical population in a homogeneous environment (Figure 2). The translation models described above predict, for whole cell populations, that the translation rate is dependent on differences in the concentration of ribosomes, mRNA and translation factors, concentrations that are themselves subject to stochasticity. This will also apply to translation in each individual cell. As the control of protein synthesis is closely linked to growth capacity, noise in translation can be expected to have a considerable impact on single-cell growth and fitness. This raises questions about how cells deal with stochasticity in gene expression rate and notably whether some mechanisms exist to dampen this noise or if cells can use it to their advantage.

Illustration of cell-to-cell heterogeneity: variation of a reporter mRNA level in a yeast cell population at steady-state

Figure 2
Illustration of cell-to-cell heterogeneity: variation of a reporter mRNA level in a yeast cell population at steady-state

(A). Schematic representation of a synthetic mRNA reporter chromosomally integrated into the HIS3 locus of S. cerevisiae. The GFP is placed under the control of the yeast PAB1 promoter and a synthetic 5′-UTR. It is fused to a sequence coding for 24 repeats of the phage MS2 aptamer (MS2SL). (B). A FISH experiment was performed on an exponentially growing yeast strain expressing the reporter mRNA with a set of 48 Stellaris® RNA FISH probes labelled with Quasar670 that hybridize to the MS2 sequence. Section (B) shows a representative image of the experiment. The top panel is a maximum z-projection filtered to remove background and to sharpen the signal using FISH Quant software [38]. The numbers to the right side of the cell represent the number of spots (i.e., single mRNA molecules) detected using FISH Quant. The bottom right panel is a magnified image of one of the cells with each spot detected by FISH Quant circled. The bottom left panel is showing the corresponding bright field image. The nucleus is outlined by a broken yellow line.

Figure 2
Illustration of cell-to-cell heterogeneity: variation of a reporter mRNA level in a yeast cell population at steady-state

(A). Schematic representation of a synthetic mRNA reporter chromosomally integrated into the HIS3 locus of S. cerevisiae. The GFP is placed under the control of the yeast PAB1 promoter and a synthetic 5′-UTR. It is fused to a sequence coding for 24 repeats of the phage MS2 aptamer (MS2SL). (B). A FISH experiment was performed on an exponentially growing yeast strain expressing the reporter mRNA with a set of 48 Stellaris® RNA FISH probes labelled with Quasar670 that hybridize to the MS2 sequence. Section (B) shows a representative image of the experiment. The top panel is a maximum z-projection filtered to remove background and to sharpen the signal using FISH Quant software [38]. The numbers to the right side of the cell represent the number of spots (i.e., single mRNA molecules) detected using FISH Quant. The bottom right panel is a magnified image of one of the cells with each spot detected by FISH Quant circled. The bottom left panel is showing the corresponding bright field image. The nucleus is outlined by a broken yellow line.

Living organisms have evolved diverse ways to control and dampen noise [14]. This is true, for example during multicellular organism development, when stochasticity is a threat for the precise spatio-temporal execution of the specific gene expression programmes it relies on [15]. On the other hand, mechanisms to reduce noise in gene expression have a significant energy cost for cells which, in turn, can constrain their evolutionary potential [16]. In other words, it is impossible to completely avoid stochasticity and cells have to find a good balance to faithfully express gene programmes while allowing a certain amount of stochasticity so that the energy costs remain reasonable. In some situations, cells can even exploit stochasticity to their advantage. Indeed, stochasticity is a critical source of non-genetic cellular diversity, from microbes to mammals and is highly involved in cellular decision making [14]. For instance, although noise is tightly controlled and dampened in most developmental processes, it is sometimes required for specific development mechanisms such as the patterning of the fruit fly's eye [17]. In addition, by creating phenotypic diversity for otherwise genetically identical cells, stochasticity is crucial for optimized resource utilization and adaptation of micro-organisms to stressful fluctuating environments [14,18]. In fact, it allows organisms to generate variant subpopulations of cells that often have a decreased fitness in their current environment but may be better adapted to a change in that environment. This is known as a ‘bet-hedging’ strategy and is widespread among micro-organisms that manifest mechanisms such as bacterial persistence, sporulation and competence [14,18]. Bacterial persistence, for example, is a process in which a subset of the population spontaneously stops growing and enters a dormant state in the absence of any environmental stress [19]. If stress arises, those non-growing persistent cells can survive, unlike the cells that are still growing. Since the persistent cells can revert slowly to normal growth, this mechanism guards against population extinction. In the yeast S. cerevisiae, Blake et al. [20] showed that cells expressing a protein conferring resistance to the antibiotic Zeocin from a high-noise version of a promoter were more able to survive and propagate after exposure to high levels of the antibiotic compared with those expressing this protein from the low-noise version of the promoter. Conversely, this increased cell-to-cell heterogeneity was less advantageous at low levels of antibiotic. Bet-hedging has also been observed in yeast in the context of carbon-source choice and consumption [21]. Finally, in a high-throughput study in S. cerevisiae, Newman et al. [22] showed that genes encoding proteins responding to environmental fluctuations, such as proteins involved in amino-acid biosynthesis and stress-responses, tend to display high levels of noise.

How does translation contribute to total gene expression stochasticity?

Despite the potentially strong impact of stochasticity in translation on cell survival, this aspect of gene expression noise has barely been investigated. Mathematical modelling of prokaryotic gene expression has revealed that for two genes expressed at the same resulting average protein abundance, the one with the higher translational efficiency and lower mRNA abundance is predicted to display greater fluctuations in protein concentration than the gene with the lower translational efficiency and higher mRNA abundance [2325]. This has been confirmed experimentally by Ozbudak et al. [26] in Bacillus subtilis where a decrease in translation efficiency of the reporter efficiently reduced total noise strength (variance divided by the mean, one of the ways used to express and analyse noise). The noise strength was found to be less dependent on variations in the transcription rate than on variations in translation rate.

Using a similar approach, Blake et al. [27] also reported an increase in noise strength when increasing translation efficiency in the eukaryote S. cerevisiae. However, they showed that this effect was more pronounced when coupled to a noisy transcriptional state. More importantly, in contrast with observations made in prokaryotes, they demonstrated that tuning transcription alone could have considerable effects on noise strength. Stochastic simulations suggested that pulsatile mRNA production was required to reproduce their system behaviour. Using S. cerevisiae as a model organism, others have found that noise is dependent on the type of promoter and not absolutely dependent on the rate of expression [23,24,2628]. It seems that slow promoter activation due to chromatin remodelling has an important role in generating stochasticity in eukaryotic gene expression. It is thought that this is one of the key differences between eukaryotes and prokaryotes, in that transition rates between on and off states of promoters in prokaryotes are assumed to be comparatively fast [25,28].

If we are to fully understand translational control at the single cell level, we need to be able to quantify the copy numbers of both mRNA and protein molecules. The degree of correlation between these two copy numbers has been questioned in at least one study [29], but generally we expect the relationship to be governed by dynamic processes of synthesis and degradation that can be modelled. RNA FISH (fluorescence in situ hybridization) and bacteriophage MS2/bacteriophage PP7 tagging systems allow direct visualization and quantification of single mRNA molecules [30,31] (Figure 2). Moreover, using PP7 tagging and widefield microscopy, it is possible to directly visualize a transcription start site and analyse transcription kinetics in single cells [32]. Different techniques exist to measure single-cell protein levels [33], with the most common and user-friendly being flow cytometry. Fluorescence microscopy has also been widely used to quantify translation at the single-cell level [34]. However, the sensitivity of these techniques does not allow the quantification of individual proteins. Another approach to monitoring translation without measuring protein production [35] detects the first round of translation by double-tagging mRNA with MS2 and PP7 systems.

In conclusion, there is still considerable uncertainty about the contributions of translation to noise generation during the gene expression process. The translation machinery is potentially a major ‘noise generator’ that will drive fluctuations in the abundance of all proteins in the cell. It has been suggested that under- or overexpression of proteins involved in multi-subunit complexes is more likely to result in a reduced growth rate or viability in yeast [36] and that, consequently, those proteins must have been selected for low levels of noise [37]. Genome-wide single-cell proteomic analyses [22] suggest that proteins involved in translation initiation manifest lower-than-average stochasticity. However, the total noise that may result from the addition of the noise of multiple factors in the pathway has yet to be assessed. In this context, it is interesting to note that a number of the strongly rate controlling factors in yeast are encoded by multiple alleles.

Future work will need to focus on defining the roles of mRNA translation, and of the translation machinery, on gene expression noise. This will undoubtedly provide new insight into important relationships between translational control and mechanisms influencing evolution and fitness.

Funding

This work was supported by the Biotechnology and Biological Sciences Research Council [grant numbers BB/1020535/1 and BB/1008349/1]; and the Biotechnology and Biological Sciences Research Council/The Engineering and Physical Sciences Research Council [grant number BB/M017982/1].

Abbreviations

     
  • eIF4A

    eukaryotic initiation factor 4A

  •  
  • FISH

    fluorescence in situ hybridization

  •  
  • MS2/PP7

    derived from MS2/PP7 bacteriophage

  •  
  • RJ

    response coefficient

Translation UK 2015: Held at the University of Aberdeen, U.K., 7–9 July 2015.

References

References
1
Cornish-Bowden
A.
Fundamentals of Enzyme Kinetics
2012
Wiley-Blackwell
2
Fell
D.A.
Metabolic control analysis: a survey of its theoretical and experimental development
Biochem. J.
1992
, vol. 
286
 (pg. 
313
-
330
)
[PubMed]
3
Fell
D.A.
Understanding the control of metabolism
Frontiers in Metabolism
1997
London
Portland Press
4
Hoops
S.
Sahle
S.
Gauges
R.
Lee
C.
Pahle
J.
Simus
N.
Singhal
M.
Xu
L.
Mendes
P.
Kummer
U.
COPASI–a COmplex pAthway sImulator
Bioinformatics
2006
, vol. 
22
 (pg. 
3067
-
3074
)
[PubMed]
5
Garrick
M.D.
The kinetics of the translation of messenger RNA into protein
J. Theor. Biol.
1967
, vol. 
17
 (pg. 
19
-
30
)
[PubMed]
6
von der Haar
T.
Mathematical and computational modelling of ribosomal movement and protein synthesis: an overview
Comput. Struct. Biotechnol. J.
2012
, vol. 
1
 pg. 
e201204002
 
[PubMed]
7
Ingolia
N.T.
Ghaemmaghami
S.
Newman
J.R.
Weissman
J.S.
Genome-wide analysis in vivo of translation with nucleotide resolution using ribosome profiling
Science
2009
, vol. 
324
 (pg. 
218
-
223
)
[PubMed]
8
Tuller
T.
Carmi
A.
Vestsigian
K.
Navon
S.
Dorfan
Y.
Zaborske
J.
Pan
T.
Dahan
O.
Furman
I.
Pilpel
Y.
An evolutionarily conserved mechanism for controlling the efficiency of protein translation
Cell
2010
, vol. 
141
 (pg. 
344
-
354
)
[PubMed]
9
Chu
D.
von der Haar
T.
The architecture of eukaryotic translation
Nucleic Acids Res.
2012
, vol. 
40
 (pg. 
10098
-
10106
)
[PubMed]
10
Brackley
C.A.
Romano
M.C.
Thiel
M.
The dynamics of supply and demand in mRNA translation
PLoS Comput. Biol.
2011
, vol. 
7
 pg. 
e1002203
 
[PubMed]
11
Firczuk
H.
Kannambath
S.
Pahle
J.
Claydon
A.
Beynon
R.
Duncan
J.
Westerhoff
H.
Mendes
P.
McCarthy
J.E.
An in vivo control map for the eukaryotic mRNA translation machinery
Mol. Syst. Biol.
2013
, vol. 
9
 pg. 
635
 
[PubMed]
12
Selimkhanov
J.
Hasty
J.
Tsimring
L.S.
Recent advances in single-cell studies of gene regulation
Curr. Opin. Biotechnol.
2012
, vol. 
23
 (pg. 
34
-
40
)
[PubMed]
13
Tsioris
K.
Torres
A.J.
Douce
T.B.
Love
J.C.
A new toolbox for assessing single cells
Ann. Rev. Chem. Biomol. Eng.
2014
, vol. 
5
 (pg. 
455
-
477
)
14
Balazsi
G.
van Oudenaarden
A.
Collins
J.J.
Cellular decision making and biological noise: from microbes to mammals
Cell
2011
, vol. 
144
 (pg. 
910
-
925
)
[PubMed]
15
Arias
A.M.
Hayward
P.
Filtering transcriptional noise during development: concepts and mechanisms
Nat. Rev. Genet.
2006
, vol. 
7
 (pg. 
34
-
44
)
[PubMed]
16
Lestas
I.
Vinnicombe
G.
Paulsson
J.
Fundamental limits on the suppression of molecular fluctuations
Nature
2010
, vol. 
467
 (pg. 
174
-
178
)
[PubMed]
17
Wernet
M.F.
Mazzoni
E.O.
Celik
A.
Duncan
D.M.
Duncan
I.
Desplan
C.
Stochastic spineless expression creates the retinal mosaic for colour vision
Nature
2006
, vol. 
440
 (pg. 
174
-
180
)
[PubMed]
18
Fraser
D.
Kaern
M.
A chance at survival: gene expression noise and phenotypic diversification strategies
Mol. Microbiol.
2009
, vol. 
71
 (pg. 
1333
-
1340
)
[PubMed]
19
Balaban
N.Q.
Merrin
J.
Chait
R.
Kowalik
L.
Leibler
S.
Bacterial persistence as a phenotypic switch
Science
2004
, vol. 
305
 (pg. 
1622
-
1625
)
[PubMed]
20
Blake
W.J.
Balazsi
G.
Kohanski
M.A.
Isaacs
F.J.
Murphy
K.F.
Kuang
Y.
Cantor
C.R.
Walt
D.R.
Collins
J.J.
Phenotypic consequences of promoter-mediated transcriptional noise
Mol. Cell
2006
, vol. 
24
 (pg. 
853
-
865
)
[PubMed]
21
Venturelli
O.S.
Zuleta
I.
Murray
R.M.
El-Samad
H.
Population diversification in a yeast metabolic program promotes anticipation of environmental shifts
PLoS Biol.
2015
, vol. 
13
 pg. 
e1002042
 
[PubMed]
22
Newman
J.R.
Ghaemmaghami
S.
Ihmels
J.
Breslow
D.K.
Noble
M.
DeRisi
J.L.
Weissman
J.S.
Single-cell proteomic analysis of S. cerevisiae reveals the architecture of biological noise
Nature
2006
, vol. 
441
 (pg. 
840
-
846
)
23
Kierzek
A.M.
Zaim
J.
Zielenkiewicz
P.
The effect of transcription and translation initiation frequencies on the stochastic fluctuations in prokaryotic gene expression
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
8165
-
8172
)
[PubMed]
24
Swain
P.S.
Elowitz
M.B.
Siggia
E.D.
Intrinsic and extrinsic contributions to stochasticity in gene expression
Proc. Natl. Acad. Sci. U.S.A.
2002
, vol. 
99
 (pg. 
12795
-
12800
)
[PubMed]
25
Thattai
M.
van Oudenaarden
A.
Intrinsic noise in gene regulatory networks
Proc. Natl. Acad. Sci. U.S.A.
2001
, vol. 
98
 (pg. 
8614
-
8619
)
[PubMed]
26
Ozbudak
E.M.
Thattai
M.
Kurtser
I.
Grossman
A.D.
van Oudenaarden
A.
Regulation of noise in the expression of a single gene
Nat. Genet.
2002
, vol. 
31
 (pg. 
69
-
73
)
[PubMed]
27
Blake
W.J.
Kaern
M.
Cantor
C.R.
Collins
J.J.
Noise in eukaryotic gene expression
Nature
2003
, vol. 
422
 (pg. 
633
-
637
)
[PubMed]
28
Raser
J.M.
O'Shea
E.K.
Control of stochasticity in eukaryotic gene expression
Science
2004
, vol. 
304
 (pg. 
1811
-
1814
)
[PubMed]
29
Taniguchi
Y.
Choi
P.J.
Li
G.W.
Chen
H.
Babu
M.
Hearn
J.
Emili
A.
Xie
X.S.
Quantifying E. coli proteome and transcriptome with single-molecule sensitivity in single cells
Science
2010
, vol. 
329
 (pg. 
533
-
538
)
[PubMed]
30
Raj
A.
van den Bogaard
P.
Rifkin
S.A.
van Oudenaarden
A.
Tyagi
S.
Imaging individual mRNA molecules using multiple singly labeled probes
Nat. Methods
2008
, vol. 
5
 (pg. 
877
-
879
)
[PubMed]
31
Buxbaum
A.R.
Haimovich
G.
Singer
R.H.
In the right place at the right time: visualizing and understanding mRNA localization
Nat. Rev. Mol Cell Biol.
2015
, vol. 
16
 (pg. 
95
-
109
)
[PubMed]
32
Larson
D.R.
Zenklusen
D.
Wu
B.
Chao
J.A.
Singer
R.H.
Real-time observation of transcription initiation and elongation on an endogenous yeast gene
Science
2011
, vol. 
332
 (pg. 
475
-
478
)
[PubMed]
33
Wu
M.
Singh
A.K.
Single-cell protein analysis
Curr. Opin. Biotechnol.
2012
, vol. 
23
 (pg. 
83
-
88
)
[PubMed]
34
Chao
J.A.
Yoon
Y.J.
Singer
R.H.
Imaging translation in single cells using fluorescent microscopy
Cold Spring Harb. Perspect. Biol.
2012
, vol. 
4
 pg. 
pii: a012310
 
[PubMed]
35
Halstead
J.M.
Lionnet
T.
Wilbertz
J.H.
Wippich
F.
Ephrussi
A.
Singer
R.H.
Chao
J.A.
Translation. An RNA biosensor for imaging the first round of translation from single cells to living animals
Science
2015
, vol. 
347
 (pg. 
1367
-
1671
)
[PubMed]
36
Papp
B.
Pal
C.
Hurst
L.D.
Dosage sensitivity and the evolution of gene families in yeast
Nature
2003
, vol. 
424
 (pg. 
194
-
197
)
[PubMed]
37
Fraser
H.B.
Hirsh
A.E.
Giaever
G.
Kumm
J.
Eisen
M.B.
Noise minimization in eukaryotic gene expression
PLoS Biol.
2004
, vol. 
2
 pg. 
e137
 
[PubMed]
38
Mueller
F.
Senecal
A.
Tantale
K.
Marie-Nelly
H.
Ly
N.
Collin
O.
Basyuk
E.
Bertrand
E.
Darzacq
X.
Zimmer
C.
FISH-quant: automatic counting of transcripts in 3D FISH images
Nat. Methods
2013
, vol. 
10
 (pg. 
277
-
278
)
[PubMed]

Author notes

1

Joint first authors