Analysis of muscle fiber structure in cardiac tissue

Collaboration

Since I am PhD student in the lab of Eberhard Bodenschatz at the Max Planck Institute for Dynamics and Self-Organization, working together with the lab of Alexander Ecker, this project is a collaboration between the two labs.

Motivation

This project focuses on implementing and comparing three computational methods for analyzing sarcomere structure in cardiac muscle. Cardiac muscle consists of thousands of interconnected cardiac muscle cells. In every cardiac muscle cell, you can find lengthwise repeating sarcomeres. They are the smallest contractile units responsible for generating force and driving the rhythmic beating of the heart. In healthy adult heart muscle tissue, sarcomeres are uniformly aligned, ensuring efficient contraction and relaxation. In contrast, heart failure is characterized by pronounced sarcomere and myofibril disarray. Therefore accurate quantification of sarcomere organization and dynamics is essential for understanding heart disease mechanisms and evaluating potential therapies.

Project

While several tools already exist to analyze sarcomere structure from fluorescent images, there is still room for innovation. We hypothesize that a physics-based pattern analysis method using Fourier analysis( based on: paper), could compete with or outperform current methods.

The goal of this project is to implement, compare, and improve methods for sarcomere analysis. You will work with a wavelet-based method (paper), an AI-based approach (paper), and the Fourier-based method (based on: paper). Most of the Python code for these methods is already available, so the focus will be on reproducing, comparing and improving the approaches.

The minimal goal of the project is to implement all three methods as described in their respective publications, document them, and compare their performance on a common dataset. A further goal is to refine the Fourier-based method to match or exceed the performance of the wavelet- and AI-based approaches.

Thesis

In this thesis, we will start by addressing the following question:

• How do existing methods and a Fourier-based approach compare in quantifying sarcomere structure in cardiomyocytes?

Building on this, we will explore different strategies to improve the performance of the Fourier-based method. This could include addressing questions such as:

• How do different approaches for muscle fiber detection (Gaussian blurring, edge detection, AI-based segmentation) affect the analysis of sarcomere structure?
• How do smoothing techniques influence the analysis of sarcomere structure?
• Is it feasible to implement real-time analysis of sarcomere length in videos?

Requirements

Some basic experience in coding is recommended, along with basic knowledge of physics or mathematics. No prior knowledge of biology is necessary.

Contact

To apply please email Ina Braun stating your interest in this project and detailing your relevant skills. A part of this project could be also a lab rotation.