Causes and consequences of stochastic processes in development and disease

17 - 18 April 2023 09:00 - 17:00 The Royal Society Free Watch online
Phenotypic heterogeneity (red or green, scored using Machine Learning) in a monolayer of clonally identical mutant IMCD3 cells

Scientific discussion meeting organised by Dr Dagan Jenkins, Dr Gabriel Galea and Professor Jonathan Chubb.

Stochastic processes impact upon many areas of biology. This phenomenon has recently become a research discipline in its own right owing to technological advances in DNA sequencing, single molecule and single cell methodologies, genetic engineering and mathematical modelling. Investigating stochastic processes requires a multidisciplinary approach, and this meeting brought together scientists from disparate areas to promote dialogue between them.

The schedule of talks and speaker biographies are available below. Meeting papers will be published in a future issue of Philosophical Transactions of the Royal Society B.

 

Attending this event

This event has now taken place.

To view the programme, please scroll down and select the day on the left-hand side. Click the arrows to view the speakers and talks.

Enquiries: contact the Scientific Programmes team

Organisers

  • Dr Dagan Jenkins

    Dr Dagan Jenkins, UCL, UK

    Dr Jenkins has spent 20 years working on the genetics of human birth defects, especially disorders of the skeleton and in particular ciliopathies. He set up his own lab as an MRC New Investigator and is currently Director of a Wellcome Trust-funded Collaborative Award in Science which focuses on missense mutations in intraflagellar transport (IFT) proteins. Using a combination of gene-editing, affinity proteomics and phenomics his laboratory, together with collaborators in London and Germany, focuses on fundamental genetic principles in relation to genetic threshold effects and developmental pleiotropy. His work has suggested an important role for stochastic processes in determining disease severity.

  • Dr Gabriel Galea

    Dr Gabriel Galea, UCL, UK

    Dr Galea is a principal research fellow at the UCL GOS Institute of Child Health leading research which aims to improve prediction, prevention and patient outcomes for individuals affected by Neural Tube Defects such as spina bifida. His previous work identified mechanisms by which embryonic cells achieve neural tube morphogenesis through coordinated force-generating behaviours such as apical constriction enhanced by Wnt/Planar cell polarity signalling. The group’s recent interests have focused on mechanisms by which stochastically acquired somatic mutations in mosaic subsets of embryonic cells can disproportionately impair neural tube development.

  • blank avatar

    Dr Jonathan Chubb, University College London, UK

Schedule

Chair

Dr Gabriel Galea

Dr Gabriel Galea, UCL, UK

09:00-09:05 Introduction
09:05-09:30 Dynamic control of cell state transitions

Cell state transitions are a fundamental backbone of biological systems. Cells often need to change properties and acquire new identities: whether it is a pluripotent cell differentiating towards a given germ layer in the early embryo, an adult stem cell dividing and specialising to replace damaged cells and maintain tissue function, or a cell becoming migratory and leaving its niche. Dysregulation of cell state transitions is also associated with various diseases, from developmental disorders to cancer. Often the transition between distinct cell states is abrupt, with sharp boundaries separating one state from another. However, it remains unclear how biological systems generate such boundaries, and how sharp transition dynamics can arise from continuous transcriptional changes. Dr Mulas will discuss the evidence for a role of differential protein translation in the control of cell state transitions dynamics both in embryonic development and in adult homeostatic tissues.

Dr Carla Mulas, Altos Labs, Cambridge Institute, UK

Dr Carla Mulas, Altos Labs, Cambridge Institute, UK

09:30-09:45 Discussion
09:45-10:15 The cardiac cell atlas: the developing and adult heart one cell at a time

Single-cell resolution profiling of gene expression and chromatin accessibility are powerful tools to study cellular diversity, interactivity and trajectories in health, development and disease. The additional dimension of spatial transcriptomics allows us map cell types defined in single-cell data to their native tissue microarchitecture and resolve niches of interacting cells. This presentation will showcase a series of biological insights gained during the construction of adult and foetal cardiac cells atlases using these modalities.

Dr James Cranley, Wellcome Sanger Institute, UK

Dr James Cranley, Wellcome Sanger Institute, UK

10:15-10:30 Discussion
10:30-11:00 Break
11:00-11:30 Multimodal perception links cellular state to decision making in single cells

Contextual decision making by individual cells in a collective is a hallmark of multicellular systems. To achieve context-aware behaviour, individual cells must integrate the input they receive from growth factors with the complexity of information of their physicochemical state and their microenvironment (cellular state). Cells perceive this through intracellular signalling networks, but individual signalling nodes are thought to have low information processing capacity, reflected in their heterogenous responses to growth factor stimulation. Dr Kramer instead hypothesised that heterogenous responses of individual signalling nodes do not reflect low information processing capacity but reflect adaptive, cellular state-dependent, information processing. When considered holistically, as a multimodal percept, activation across the whole network could provide single cells with sufficient information to enable accurate contextual decision making. To test this, Dr Kramer and his group utilised EGF stimulation, combined with multiplexed (40-plex), imaging-based quantification to generate comprehensive single-cell data mapping the signalling responses across the EGFR pathway and cellular states in millions of individual cells. They find that signalling nodes in networks indeed display deterministic and adaptive cellular state-dependent information processing, which leads to heterogeneous growth factor responses and enables nodes to capture non-redundant information about the cellular state. Collectively, as a multimodal percept, activation across the network reflects the diversity of cellular states in cell populations. They further find that this multimodal percept accurately reflects varying input concentrations of EGF. Lastly, they find that multimodal perception links the cellular state to the heterogenous decision of single cells to re-enter the cell cycle or stay quiescent after exposure of EGF. 

Dr Bernhard Kramer, University of Zurich, Switzerland

Dr Bernhard Kramer, University of Zurich, Switzerland

11:30-11:45 Discussion
11:45-12:15 Modelling the dynamics of development with time-resolved transcriptomics

Recent advances in single-cell transcriptome assays, such as RNA velocity and metabolic labelling, provide temporal information into the dynamics of gene expression. This promises to offer insight into mechanisms driving cell state transitions and developmental fate acquisition. However, current methods have experimental limitations and computational caveats that restrict their utility. Here, Rory Maizels presents an integrated pipeline for dynamical modelling of gene expression from single-cell data. He has developed an optimised, automated protocol for single-cell RNA-sequencing with metabolic labelling, which provides substantial improvements on previous protocols, enabling the collection of high-quality temporal transcriptome profiles in over ten thousand cells per experiment. Alongside this, he has developed a deep learning approach for inferring the transcriptional dynamics of a system within the latent space of a variational autoencoder model. By incorporating neighbourhood information into the learning process, Rory will demonstrate how the solution space for a cell’s dynamics can be constrained based on the distribution of its local data manifold. In doing so, he finds his method outperforms current velocity inference methods, tested in quantitative benchmarking. With the improved velocity inference, Rory will discuss how learnt dynamics can be embedded into a neural stochastic differential equation system within the latent space, allowing probabilistic simulations of cellular behaviour. Applying this to a dataset collected from neuronal differentiation, he will show how dynamical aspects, such as the decision boundaries between bifurcating fates and the time-directed correlations between key genes, can capture and provide a step towards quantitative modelling from single cell data.

Mr Rory Maizels, Briscoe Laboratory, Francis Crick Institute, UK

Mr Rory Maizels, Briscoe Laboratory, Francis Crick Institute, UK

12:15-12:30 Discussion

Chair

Cristina Pina

Dr Cristina Pina, Brunel University London, UK

13:30-14:00 Effects of cell cycle variability on stochastic gene expression in a population of cells

Many models of stochastic gene expression do not incorporate a cell cycle description. Dr Perez-Carrasco will show how this can be tackled analytically studying how mRNA fluctuations are influenced by DNA replication for a prescribed cell cycle duration stochasticity. Results show that omitting cell cycle details can introduce significant errors in the predicted mean and variance of gene expression for prokaryotic and eukaryotic organisms, reaching 25% error in the variance for mouse fibroblasts. Furthermore, he will show how population statistics differ from single cell trajectories making emphasis on the relevance of discerning self-organisation mechanisms from simple non-linearities.

Dr Ruben Perez-Carrasco, University College London, UK

Dr Ruben Perez-Carrasco, University College London, UK

14:00-14:15 Discussion
14:15-14:45 Quantifying how post-transcriptional noise and gene copy number variation bias transcriptional parameter inference from mRNA distributions

Transcriptional rates are often estimated by fitting the distribution of mature mRNA numbers measured using smFISH (single molecule fluorescence in situ hybridisation) with the distribution predicted by the telegraph model of gene expression, which defines two promoter states of activity and inactivity. However, fluctuations in mature mRNA numbers are strongly affected by processes downstream of transcription. In addition, the telegraph model assumes one gene copy but in experiments, cells may have two gene copies as cells replicate their genome during the cell cycle. While it is often presumed that post-transcriptional noise and gene copy number variation affect transcriptional parameter estimation, the size of the error introduced remains unclear. To address this issue, here we measure both mature and nascent mRNA distributions of GAL10 in yeast cells using smFISH and classify each cell according to its cell cycle phase. We infer transcriptional parameters from mature and nascent mRNA distributions, with and without accounting for cell cycle phase and compare the results to live-cell transcription measurements of the same gene. We find that: (i) correcting for cell cycle dynamics decreases the promoter switching rates and the initiation rate, and increases the fraction of time spent in the active state, as well as the burst size; (ii) additional correction for post-transcriptional noise leads to further increases in the burst size and to a large reduction in the errors in parameter estimation. Furthermore, we outline how to correctly adjust for measurement noise in smFISH due to uncertainty in transcription site localisation when introns cannot be labelled. Simulations with parameters estimated from nascent smFISH data, which is corrected for cell cycle phases and measurement noise, leads to autocorrelation functions that agree with those obtained from live-cell imaging.

Professor Ramon Grima, University of Edinburgh, UK

Professor Ramon Grima, University of Edinburgh, UK

14:45-15:00 Discussion
15:00-15:30 Break
15:30-16:00 Poster flash talk session
16:00-16:15 Discussion
16:15-17:00 Poster session

Chair

blank avatar

Dr Jonathan Chubb, University College London, UK

09:00-09:30 The role of epigenomics, stochasticity and entropy in cancer development

Recent developments in genomics indicate that epigenetic alterations may play an important causal role in predisposing normal cells to neoplastic transformation. Professor Teschendorff will begin by presenting data that support an epigenetic stem cell model of oncogenesis, according to which DNA methylation (DNAm) changes that accrue in normal cells as a function of age and other cancer risk factors, lead to differentiation defects and increased aberrant plasticity, promoting cancer development. In the context of two cancers (cervical and breast), he will demonstrate how specific DNAm changes in normal tissue can discriminate those tissues at future risk of neoplastic transformation from those that remain healthy. In the earliest stages these DNAm changes are inherently stochastic across individuals but exhibit selection and convergence upon neoplastic transformation. He will also describe how DNAm changes can be used to measure mitotic-age of tissues, and that DNAm-based mitotic clocks could be used for cancer risk prediction. In the last part of the talk, Professor Teschendorff will describe the group's efforts to quantify cancer risk at the resolution of single-cells via a bottom-up network model that quantifies dedifferentiation from single-cell RNA-Seq data in terms of diffusion entropy. Specifically, he will show how in the context of a multi-stage model of oesophageal cancer development, diffusion entropy identifies the precancerous cells that undergo selection during cancer progression. Mounting evidence points towards underlying epigenetic alterations being better cancer risk markers than genetic mutations or copy-number changes. He will end by discussing the implications of these findings for developing cancer risk prediction strategies.

Professor Andrew Teschendorff, CAS Key Lab of Computational Biology, Shanghai Institute for Nutrition and Health, Chinese Academy of Sciences, China

Professor Andrew Teschendorff, CAS Key Lab of Computational Biology, Shanghai Institute for Nutrition and Health, Chinese Academy of Sciences, China

09:30-09:45 Discussion
09:45-10:15 Insight into disease susceptibility and stochastic noise in syndromic disease from a multi-modal mouse model

Cilia are membrane-encased organelles present on the surface of most cells that rely upon intraflagellar transport (IFT) for their formation and function. Missense mutations in IFT genes cause skeletal, limb, eye and renal disease, and the group modelled several Ift80 mutations in mice. They find that some tissues are more susceptible to disease than others, and they describe incomplete penetrance in the absence of genetic background variation suggesting that stochastic processes determine disease susceptibility. One missense mutation within a splice acceptor affected multiple modes of gene function at the levels of RNA splicing, protein abundance and protein complex formation. This provides a means to investigate different sources of stochastic noise, and to use this biological variation to understand how mutational effects at different biological levels contribute to disease risk. A large proportion of mutations in human monogenic disease are also likely to be multimodal suggesting a generalised approach to investigate stochastic processes that could be applied to many diseases.

Dr Dagan Jenkins, UCL, UK

Dr Dagan Jenkins, UCL, UK

10:15-10:30 Discussion
10:30-11:00 Break
11:00-11:30 Contributions of transcriptional noise to leukaemia evolution

In this talk, Dr Pina will discuss the contributions of epigenetic heterogeneity and transcriptional noise in initiation and maintenance of Acute Myeloid Leukaemia. She will focus on KAT2A, a histone acetyltransferase (HAT) previously shown to regulate transcriptional noise levels in mammalian stem cell systems, and discuss the group's recent observations that Kat2a loss can both promote pre-leukaemia and impede leukaemia maintenance, with similar effects on variability on gene transcription and consequent cell diversification. She will discuss KAT2A roles through specific HAT complex integration, target programmes, and remodelling of gene regulatory networks, and integrate the findings with the transcriptional consequences of other epigenetic complexes previously implicated in regulation of transcriptional noise and progression of leukaemia. Drawing from stem cell and leukaemia models, she will discuss gene-specific responses to transcriptional noise modulation, and suggest potential contextual contributions to functional outcomes and putative value as markers of leukaemia progression and therapeutic response.

Dr Cristina Pina, Brunel University London, UK

Dr Cristina Pina, Brunel University London, UK

11:30-11:45 Discussion
11:45-12:15 Control of celluar noise by subcellular compartments

Chemical reactions serve as basic units for cellular information processing and control. However, reactions in cells are noisy, leading to substantial variability in the molecular constitution of genetically identical cells. How cells deal with such noise has been important question in quantitative biology. In this talk, Dr Zechner will discuss how the mesoscale organisation of molecules into condensates can be used by cells to suppress cellular noise. Using concepts from statistical physics and control theory, he will show that condensates mediate negative feedback, which maintains concentration levels at precise set points in the presence of noise. He will present experimental single-cell data from synthetic and endogenous systems supporting this prediction.

Dr Christoph Zechner, Max Planck Institute of Molecular Cell Biology and Genetics, Germany

Dr Christoph Zechner, Max Planck Institute of Molecular Cell Biology and Genetics, Germany

12:30-13:30 Discussion

Chair

Dr Dagan Jenkins

Dr Dagan Jenkins, UCL, UK

13:30-14:00 Stochastic cell fate specification in the fly eye

Stochastic mechanisms diversify cell fates during development. How cells randomly choose between two or more fates remains poorly understood. In the Drosophila eye, the random mosaic of two R7 photoreceptor subtypes is determined by expression of the transcription factor Spineless (Ss). The group investigated how cis-regulatory elements and trans factors regulate nascent transcriptional activity and chromatin compaction at the ss gene locus during R7 development. The ss locus is in a compact state in undifferentiated cells. An early enhancer drives transcription in all R7 precursors, and the locus opens. In differentiating cells, transcription ceases and the ss locus stochastically remains open or compacts. In SsON R7s, ss is open and competent for activation by a late enhancer, whereas in SsOFF R7s, ss is compact, and repression prevents expression. These results suggest that a temporally dynamic antagonism, in which transcription drives large-scale decompaction and then compaction represses transcription, controls stochastic fate specification.

Associate Professor Robert J Johnston Jr, Johns Hopkins University, USA

Associate Professor Robert J Johnston Jr, Johns Hopkins University, USA

14:00-14:15 Discussion
14:15-14:45 A functional role for noise in plant development

Gene expression in individual cells can be surprisingly noisy. In unicellular organisms this noise can be functional. For example, it can allow a fraction of the population to survive a sudden environmental stress. The role of gene expression noise in multicellular organisms is less clear. In this talk, Dr Locke will show how new techniques are revealing an unexpected level of variability in gene expression between and within genetically identical plants. He will discuss evidence that transcriptional noise can act as a mechanism for generating functional phenotypic diversity in plants.

Dr James Locke, University of Cambridge, UK

Dr James Locke, University of Cambridge, UK

14:45-15:00 Discussion
15:00-15:30 Break
15:30-16:00 No precision without plasticity

Development from fertilised egg to functioning multi-cellular organism requires precision. Precision can only be achieved through plasticity. Plasticity is conferred in part by stochastic variation, which is inherently present in biological systems. Gene expression levels fluctuate at multiple stages: transcription, alternative splicing, translation, and turnover. Evolving multicellularity and cellular specialisation have exploited the small differences in gene expression to confer distinct functions on selected individual cells. This initial differentiation sets in motion regulatory interactions allowing the non-selected cells to acquire new functions along the spatiotemporal developmental trajectory. Many other components of the differentiation process are stochastic. The dynamic binding of transcriptional regulatory assemblies to DNA exhibit randomness. X-inactivation is normally random, non-random inactivation is generally a sign of mutational perturbation. Correct neural wiring arises through random connections, followed by maintenance only for functional links. This mechanism is well illustrated for the retina. In immune system development both antibody maturation in B cells and the emergence of a balanced population of T-cells begins through stochastic trial and error followed by functional selection. Multiple manifestations of disease arise when the switch from random fluctuation to established pattern selection fails. A wide spectrum of mostly deleterious DNA mutations through stochastic fluctuation in environmental factors: such as levels of radiation or presence of pollutants. The phenotypic outcome of mutations is dependent on variability across genomic functions and the environment. Variable environmental factors also determine which mutations prove advantageous. To ensure population survival, bet-hedging stochastic strategies are often deployed.

Professor Veronica Van Heyningen FRS, University of Edinburgh, UK

Professor Veronica Van Heyningen FRS, University of Edinburgh, UK

16:00-16:15 Discussion
16:15-17:00 Panel discussion