Publications

The functional relevance of an organoid is dependent on the differentiation, morphology, cell arrangement and biophysical properties, which collectively define the state of an organoid. For an organoid culture, an individual organoid or the cells that compose it, these state variables can be characterised, most easily by transcriptomics and by high-content image analysis. Their states can be compared to their in vivo counterparts. Current evidence suggests that organoids explore a wider state space than organs in vivo due to the lack of niche signalling and the variability of boundary conditions in vitro. Using data-driven state inference and in silico modelling, phase diagrams can be constructed to systematically sort organoids along biochemical or biophysical axes. These phase diagrams allow us to identify control strategies to modulate organoid state. To do so, the biochemical and biophysical environment, as well as the cells that seed organoids, can be manipulated.

We investigate a class of DNA mixture deconvolution algorithms based on variational inference, and we show that this can significantly reduce computational runtimes with little or no effect on the accuracy and precision of the result. In particular, we consider Stein Variational Gradient Descent (SVGD) and Variational Inference (VI) with an evidence lower-bound objective. Both provide alternatives to the commonly used Markov-Chain Monte-Carlo methods for estimating the model posterior in Bayesian probabilistic genotyping. We demonstrate that both SVGD and VI significantly reduce computational costs over the current state of the art. Importantly, VI does so without sacrificing precision or accuracy, presenting an overall improvement over previously published methods.

Correct nervous system development depends on the timely differentiation of progenitor cells into neurons. While the output of progenitor differentiation is well investigated at the population and clonal level, how stereotypic or variable fate decisions are during development is still more elusive. To fill this gap, we here follow the fate outcome of single neurogenic progenitors in the zebrafish retina over time using live imaging. We find that neurogenic progenitor divisions produce two daughter cells, one of deterministic and one of probabilistic fate. Interference with the deterministic branch of the lineage affects lineage progression. In contrast, interference with fate probabilities of the probabilistic branch results in a broader range of fate possibilities than in wild-type and involves the production of any neuronal cell type even at non-canonical developmental stages. Combining the interference data with stochastic modelling of fate probabilities revealed that a simple gene regulatory network is able to predict the observed fate decision probabilities during wild-type development. These findings unveil unexpected lineage flexibility that could ensure robust development of the retina and other tissues.

ATPases are a group of enzymes that can cyclically convert the free energy of ATP hydrolysis into mechanical work. GTPases are another class of enzymes that are predominantly associated with signal transduction processes, but their role in mechanotransduction is less established. It was previously shown that the binding of the GTPase Rab5 to the tethering protein EEA1 induces a large conformational change in EEA1 from a rigid, extended to a flexible, collapsed state. This entropic collapse of EEA1 gives rise to an effective force that can pull tethered membranes closer. It currently remains unclear if EEA1 can return from the collapsed to the extended conformation without the aid of chaperone proteins. Here we show that EEA1 in a bulk solution can undergo multiple flexibility transition cycles driven by the energetics of Rab5 binding and unbinding as well as GTP hydrolysis. Each cycle can perform up to 20kBT of mechanical work. Hence, Rab5 and EEA1 constitute a two-component molecular motor driven by the chemical energy derived from the Rab5 GTPase cycle. We conclude that tethering proteins and their small GTPase partners can have active mechanical roles in membrane trafficking.

Horizontal transfer of transposable elements (TEs) is an important mechanism contributing to genetic diversity and innovation. Bats (order Chiroptera) have repeatedly been shown to experience horizontal transfer of TEs at what appears to be a high rate compared with other mammals. We investigated the occurrence of horizontally transferred (HT) DNA transposons involving bats. We found over 200 putative HT elements within bats; 16 transposons were shared across distantly related mammalian clades, and 2 other elements were shared with a fish and two lizard species. Our results indicate that bats are a hotspot for horizontal transfer of DNA transposons. These events broadly coincide with the diversification of several bat clades, supporting the hypothesis that DNA transposon invasions have contributed to genetic diversification of bats.

The discovery of biomolecular condensates transformed our understanding of intracellular compartmentalization of molecules. To integrate interdisciplinary scientific knowledge about the function and composition of biomolecular condensates, we developed the crowdsourcing condensate database and encyclopedia ( cd-code.org ). CD-CODE is a community-editable platform, which includes a database of biomolecular condensates based on the literature, an encyclopedia of relevant scientific terms and a crowdsourcing web application. Our platform will accelerate the discovery and validation of biomolecular condensates, and facilitate efforts to understand their role in disease and as therapeutic targets.

Negative-stranded RNA viruses can establish long-term persistent infection in the form of large intracellular inclusions in the human host and cause chronic diseases. Here, we uncover how cellular stress disrupts the metastable host-virus equilibrium in persistent infection and induces viral replication in a culture model of mumps virus. Using a combination of cell biology, whole-cell proteomics, and cryo-electron tomography, we show that persistent viral replication factories are dynamic condensates and identify the largely disordered viral phosphoprotein as a driver of their assembly. Upon stress, increased phosphorylation of the phosphoprotein at its interaction interface with the viral polymerase coincides with the formation of a stable replication complex. By obtaining atomic models for the authentic mumps virus nucleocapsid, we elucidate a concomitant conformational change that exposes the viral genome to its replication machinery. These events constitute a stress-mediated switch within viral condensates that provide an environment to support upregulation of viral replication.

Physical mechanisms of phase separation in living systems play key physiological roles and have recently been the focus of intensive studies. The strongly heterogeneous nature of such phenomena poses difficult modeling challenges that require going beyond mean-field approaches based on postulating a free energy landscape. The pathway we take here is to calculate the partition function starting from microscopic interactions by means of cavity methods, based on a tree approximation for the interaction graph. We illustrate them on the binary case and then apply them successfully to ternary systems, in which simpler one-factor approximations are proved inadequate. We demonstrate the agreement with lattice simulations and contrast our theory with coacervation experiments of associative de-mixing of nucleotides and poly-lysine. Different types of evidence are provided to support cavity methods as ideal tools for modeling biomolecular condensation, giving an optimal balance between the consideration of spatial aspects and fast computational results.

Cortical projection neurons polarize and form an axon while migrating radially. Even though these dynamic processes are closely interwoven, they are regulated separately-the neurons terminate their migration when reaching their destination, the cortical plate, but continue to grow their axons. Here, we show that in rodents, the centrosome distinguishes these processes. Newly developed molecular tools modulating centrosomal microtubule nucleation combined with in vivo imaging uncovered that dysregulation of centrosomal microtubule nucleation abrogated radial migration without affecting axon formation. Tightly regulated centrosomal microtubule nucleation was required for periodic formation of the cytoplasmic dilation at the leading process, which is essential for radial migration. The microtubule nucleating factor γ-tubulin decreased at neuronal centrosomes during the migratory phase. As distinct microtubule networks drive neuronal polarization and radial migration, this provides insight into how neuronal migratory defects occur without largely affecting axonal tracts in human developmental cortical dysgeneses, caused by mutations in γ-tubulin.

Despite the increasing use of genomic sequencing in clinical practice, the interpretation of rare genetic variants remains challenging even in well-studied disease genes, resulting in many patients with Variants of Uncertain Significance (VUSs). Computational Variant Effect Predictors (VEPs) provide valuable evidence in variant assessment, but they are prone to misclassifying benign variants, contributing to false positives. Here, we develop Deciphering Mutations in Actionable Genes (DeMAG), a supervised classifier for missense variants trained using extensive diagnostic data available in 59 actionable disease genes (American College of Medical Genetics and Genomics Secondary Findings v2.0, ACMG SF v2.0). DeMAG improves performance over existing VEPs by reaching balanced specificity (82%) and sensitivity (94%) on clinical data, and includes a novel epistatic feature, the 'partners score', which leverages evolutionary and structural partnerships of residues. The 'partners score' provides a general framework for modeling epistatic interactions, integrating both clinical and functional information. We provide our tool and predictions for all missense variants in 316 clinically actionable disease genes (demag.org) to facilitate the interpretation of variants and improve clinical decision-making.