Microorganisms often live in dense, surface-attached communities called biofilms. These biofilms exist in diverse environments, from the inside of your gut and the surfaces of your teeth, to the roots of your house plants and the inside of your coffee machine. These biofilms, composed of various bacteria, fungi, and viruses encased in a self-produced extracellular matrix, exhibit complex behaviours. They also play a dual role in human health and industry, contributing to persistent infections and antibiotic resistance while being crucial for digestion and industrial applications like wastewater treatment. Understanding biofilm formation and function is essential for decreasing their detrimental effects and increasing their potential in biotechnology.
In this thesis, I use individual-based modelling of spherocylindrical particles to learn something about the effects of spatial structure on their mechanical and social interactions.
In chapter 2 we explore the aggregation dynamics of blue-light switchable adhesive E. coli in solution. We aim to understand experimental results that bacteria aggregate more and formed bigger clusters under pulsating light. We simulate a system of particles undergoing Brownian motion, where the cell-cell adhesion can be periodically turned on and off and compare and match our simulations to the experimental data. We show how tuning the light off-period to the decay time of the adhesion leads to increased clustering. We conclude that partial disassembly of the aggregates leads to more effective clustering. In addition, our co-authors show that this increased clustering leads to increased biofilm formation in a laboratory setting. Moreover, it can be used to increase productivity in a bioreactor.
We use what we learnt about cell-cell interactions to simulate growing surface attached systems in chapter 3. We motivate some choices about the interactions between cells and the interaction with the surface. We then show how varying the strengths of these interactions can lead to different microcolony architectures.
We then use this model of growing microcolonies to study cooperator interactions in a spatially structured environment. Where the mechanical interactions occur over short distances, we also assume that metabolic interactions are close range. In chapter 4, we simulate a cross-feeding consortium in the presence of a cheater species by having particles adjust their growth rate based on the cells in their immediate environment. Using simulations and an experimental consortium, we show how the time it takes for cooperators to meet is the determining factor in whether they outcompete their cheating counterparts.
Finally, in chapter 5 we explore the patterning that cooperating particles create by mixing. We show that this cooperator mixing is mostly determined by interaction strength and is robust against variations in size and interaction symmetry. Additionally, we show that in the presence of cheaters, cooperators intermix but cheaters don’t mix with the cooperators and instead remain on the outside. Therefore, we argue that focusing on strong cooperation is a great strategy for cheater exclusion.
Have fun!@en

Interactions are essential to life. At the cellular level, the most fundamental processes rely on interactions between DNA, RNA, and proteins. Given that the structure and function of these molecules are determined by their sequence, it is ultimately the sequence that dictates the strength and specificity of these interactions. Thus, to improve our understanding of biological processes, and even predict or manipulate them, it is imperative to obtain a comprehensive and quantitative overview of the relationship between sequence and interactions. One valuable method for studying these interactions in detail is single-molecule fluorescence microscopy. However, this technique is expensive and labor-intensive, making it impractical for large sequence libraries. In this thesis, we introduce a novel approach called SPARXS: Single-molecule Parallel Analysis for Rapid eXploration of Sequence space. By integrating single-molecule fluorescence microscopy with next-generation sequencing, this platform enables the simultaneous profiling of interactions with thousands of different DNA or RNA sequences at the single-molecule level.@en

Predicting off-target cleavage sites for CRISPR/Cas nucleases on the human genome based on biophysical modelling and high-throughput biochemical profiling.

This project aims to describe crude models that work towards a mathematical model for transcription and translation. Transcription refers to the conversion of genetic code to a readable format for the cell, while translation refers to the conversion of this code into polypeptide chains using ribosomes. The main proposal for such a model is the totally asymmetric exclusion process which has already been applied to the case of translation [1]. In this report two simpler models and its mathematical abstractions are examined. These two are independent random walks on a grid Zd and the totally symmetric exclusion process on Zd where we allow a maximum of one particle per site. Since transcription and translation deal with particles being added and removed, we also consider the two models under which a source is placed in the origin for particles. Starting with independent random walkers, we find a principle called duality. This says that we can describe the independent random walks of a whole configuration by letting a single particle execute a random walk starting from x and evaluate the random walk at a later time t. Since precise quantitative behaviour is rather dificult to extract without further assumptions, we instead turn towards invariant distributions. It is proven that the Poisson measure is invariant on Zd, with and without a source in the origin for independent random walkers.

In this thesis the CRISPR-Cas9 mechanism, a promising mechanism for geneediting, is considered. Closed form expressions are derived for the probability and time to cleave or unbind for the associated Cas9 protein. The mechanism can be modelled mathematically by a birth and death process, therefore the expressions could be derived using Markov chains and semigroups. The expressions are compared to simulations and interpreted using the model of hybridization kinetics. Finally the moment generating function of the stopping time is derived for two special Markov processes, i.e. a random walk and a Brownian motion with drift. This was done using martingales.