Researchers have introduced the groundbreaking BeatProfiler, a comprehensive software that automates the analysis of heart cell function from video data. It is the first system to integrate the analysis of various heart function indicators, including contractility, calcium handling, and force output into one tool, which significantly speeds up the process and reduces the chance for errors. BeatProfiler allows researchers to differentiate between different diseases and their severity, as well as rapidly and objectively test drugs that impact heart function.The study of heart disease and testing new heart drugs has always been a complex and time-consuming process. One way to improve this is by using cellular and engineered tissue models in a dish, but the current methods for studying heart cell contraction and calcium handling are manual, error-prone, and require expensive specialized equipment. There is a clear need for a more efficient and accessible way to study heart function using artificial intelligence (AI) and machine learning. The BeatProfiler, a new tool, is aiming to address this need.
Researchers at Columbia Engineering have introduced a revolutionary tool called BeatProfiler, which is designed to quickly analyze the function of heart cells. This software addresses the challenges of analyzing heart cell function by automating the process of analyzing video data. BeatProfiler is the first system to integrate the analysis of various indicators of heart function, such as contractility, calcium handling, and force output, into one tool. This integration speeds up the analysis process and reduces the potential for errors. With the help of BeatProfiler, researchers are able to differentiate between different diseases and determine their severity, ultimately advancing the understanding of heart cell function.
The testing of drugs that impact heart function can now be done rapidly and objectively thanks to a new study published in the IEEE Open Journal of Engineering in Medicine and Biology on April 8. Gordana Vunjak-Novakovic, project leader and University Professor at Columbia, described the new tool as “truly transformative” due to its speed, comprehensiveness, automation, and compatibility with various computer platforms. The software is also open-source, making it easily accessible to investigators and clinicians.The professor of medicine (in Cardiology) at Columbia University Irving Medical Center chose not to apply for a patent for the AI software and instead made it open source for any lab to use for free. This decision was made to ensure the results of their research are widely disseminated and to gather feedback from users in academic, clinical, and commercial labs to improve the software. The project was motivated by the urgent need for quick and accurate diagnosis of heart disease, which is a focus of much of Vunjak-Novakovic’s research.The team worked for several years to develop a tool that could quickly and accurately assess heart diseases. Their main goal was to capture the function of the cardiac models they were building to study heart diseases and evaluate potential treatments. The researchers had an urgent need to assess the function of their cardiac models in real-time as they were making more cardiac tissues through innovations such as milliPillar and multiorgan tissue.models, the increased capabilities of the tissues required the researchers to develop a method to more rapidly quantify the function of cardiomyocytes (heart muscle cells) and tissues to enable studies exploring genetic cardiomyopathies, cosmic radiation, immune-mediated inflammation, and drug discovery.
Collaborators in software development, machine learning, and more
In the last year and a half, lead author Youngbin Kim and his coauthors developed a graphical user interface (GUI) on top of the code so that biomedical researchers with no coding expertise could easily analyze the data with just a few clicks. ThThis research brought together experts in various fields including software development for GUI, machine learning for computer vision technology and disease/drug classifiers, signal processing for contractile and calcium signals, engineering for translating pillar deflection on the cardiac platform to mechanical force, and user experience by lab members providing feedback for interface improvements.
The findings
The research demonstrated that BeatProfiler was able to effectively analyze cardiomyocyte function, surpassing existing tools in terms of speed – up to 50 times faster in some cases – and reliability. It detectedThe BeatProfiler tool can detect subtle changes in engineered heart tissue that other tools might not catch. According to Kim, a PhD candidate in Vunjak-Novakovic’s lab at Columbia Engineering, the analysis speed and versatility of BeatProfiler are unparalleled in cardiac research. Machine learning helps to distinguish between diseased and healthy heart cells with high accuracy and classify different cardiac drugs based on how they affect the heart. The team is now focused on expanding BeatProfiler’s capabilities for new applications in heart research.BeatProfiler is being developed to address a range of heart-related conditions and to aid in drug development. In order to make it applicable to a wide range of research inquiries, the team is conducting tests and validation across various in vitro cardiac models, including different types of engineered heart tissue models. Additionally, they are working on refining the machine-learning algorithm to broaden its application to different heart diseases and drug effects. Ultimately, the aim is to adapt BeatProfiler for pharmaceutical settings in order to accelerate the testing of numerous potential drugs simultaneously.
