M.Sc. Biophysics Student · Humboldt University of Berlin
Systems Biology | Systems Medicine | Computational Pharmacology
A reusable pipeline for analysing whole-slide images of liver tissue. Machine learning quantifies tissue damage from histology slides; an interactive viewer displays the results as composable layers over the original gigapixel images. Developing as my Master's project in the König Lab.
An interactive web-based tool for visualization of Finite Element Method (FEM) simulation results, with a focus on biomedical applications. Developed as part of DFG SPP 2311 and installable as an open-source Python package via PyPI.
CNN-based prediction of mean ribosome load (MRL) from 5′ UTR sequences, investigating how reducing polysome profiling resolution affects model performance. Trained on 280,000 sequences with 5-fold cross-validation across five resolution schemes; model interpretation via Integrated Gradients (Captum).
Interactive web application implementing the glimepiride digital twin for personalized and stratified diabetes treatment, built with Marimo and deployable via Docker. Accompanies the glimepiride PBPK/PD model and the corresponding publication in Frontiers in Pharmacology.
Physiologically based pharmacokinetic (PBPK) model of glimepiride distributed in SBML format, including liver, kidney, intestine, and whole-body submodels. Accompanied by a COMBINE archive, FAIR assessment, and reproducible simulation pipeline.
Work on whole-slide images of liver tissue in the König Lab: visualisation of the gigapixel slides, CNNs and transformers to segment the tissue and quantify damage, and simulation of drug metabolism and damage on the resulting geometry. Scope and title are still WIP.
Development of a whole-body PBPK model of glimepiride based on curated clinical data, investigating pharmacokinetic variability due to renal and hepatic impairment, bodyweight and CYP2C9 genetic variants.