TY - THES A1 - Drees, Philipp T1 - Python-based Data Analysis of High-Throughput Growth Curve Screenings N2 - When studying microbial growth, a rudimentary procedure involves the observation of growth factors. These factors have to be determined, reconstructed and evaluated. Considering large data quantities, the time it takes to prepare and evaluate these samples increases. By automating and concurrently simulating the microbial samples, time and human error can be reduced. The purpose behind this thesis is to document and develop the design of a Python-based software that connects to the output of the GrowthProfiler in such a way, that manual labor is reduced or replaced by automated processes as much as possible. In order to accomplish this, programming was used to get access to the recorded data. Additionally, applying analytical tools and procedures results into appropriate and readable conclusions for further interpretation. The technological development has described a decisive increase in the usage of tests and thus, instruments for multiple testing (batch experiments, numerous parallel runs) were engineered. The amount of testing increases, so there is a resulting need for ’batchevaluation’ of those tests. Manual plating of microorganisms and noting down the measurements is often replaced by devices like the GrowthProfiler, that can run 96 experiments simultaneously. However, high throughput experiments also require high throughput data analysis. Identifying the exponential phase and approximating the maximum µ, that is used in the exponential growth model. The mathematical approach behind this method is a separation of the measured growth curve into ’bins’ and comparing the slope within this ’bin’ to each other. When finding a distribution, which describes a better R², replacing it until no better fit is found. This ensures to receive a linear regression with the highest R², but not necessarily with the steepest slope. This attempt has been approached during the design of this software. With the programming language Python, a code has been constructed to detect all CSV s in certain computer folder. The data will be plotted, and the different calculations such as recognizing the biggest R² and deriving the growth rate, summarizing arithmetic means and standard deviation, concentrating all gathered information into an Excel sheet. This software was compared with other approaches to define automatically a regression curve fitting to that plot. After finalisation, this software is accessible online on a server and no downloads are necessary and can be embedded into the sequence chain between the GrowthProfiler and the final findings. Ultimately, other tools to provide programmatic support to microbiological analysis were experimented with, like simulation of the substrate uptake rate during simulated growth experiment via Monod-equation. KW - Simulation KW - Python KW - Notebook KW - Microbial growth Y1 - 2022 ER -