MyoFInDer: An AI-Based Tool for Myotube Fusion Index Determination

Abstract

The fusion index is a key indicator for quantifying the differentiation of a myoblast population, which is often calculated manually. In addition to being time-consuming, manual quantification is also error prone and subjective. Several software tools have been proposed for addressing these limitations but suffer from various drawbacks, including unintuitive interfaces and limited performance. In this study, we describe MyoFInDer, a Python-based program for the automated computation of the fusion index of skeletal muscle. At the core of MyoFInDer is a powerful artificial intelligence-based image segmentation model. MyoFInDer also determines the total nuclei count and the percentage of stained area and allows for manual verification and correction. MyoFInDer can reliably determine the fusion index, with a high correlation to manual counting. Compared with other tools, MyoFInDer stands out as it minimizes the interoperator variability, minimizes process time and displays the best correlation to manual counting. Therefore, it is a suitable choice for calculating fusion index in an automated way, and gives researchers access to the high performance and flexibility of a modern artificial intelligence model. As a free and open-source project, MyoFInDer can be modified or extended to meet specific needs.

Impact Statement

We present an efficient and user-friendly image analysis tool, MyoFInDer, to quantify myotube fusion index and the total number of nuclei in an automated and reliable way. MyoFInDer uses fluorescence microscopy images from both 2D cell cultures and 3D tissue cultures. It is of direct use for muscle tissue engineering and can be adapted in other fields. The software combines the best of both AI-based automated image analysis and manual correction. The use of MyoFInDer saves users time and energy compared with manual quantification and lowers the operator bias within research teams.