Fruit flies, scientifically known as Drosophila, serve as a valuable model for studying human heart-related issues like cardiac aging and cardiomyopathy. However, a bottleneck in evaluating fruit fly hearts is the manual measurement required at specific points in the heart’s cycle to analyze cardiac dynamics.
Researchers at the University of Alabama at Birmingham have found a way to streamline this analysis process by utilizing deep learning and high-speed video microscopy to assess more of the heart region during each heartbeat in fruit flies.
Girish Melkani, Ph.D., an associate professor at UAB, explains, “Our machine learning approach not only speeds up the process but also reduces human error by eliminating the need to manually mark heart walls during systolic and diastolic conditions. This allows for the analysis of multiple hearts simultaneously.”
This advancement opens up new possibilities to explore how environmental and genetic factors impact heart aging and diseases. Melkani foresees using deep learning in studying cardiac mutation models in various small animal models like zebrafish, mice, and potentially human heart models to gain insights into cardiac health and diseases.
The fruit fly model has proven to be highly beneficial in understanding the underlying causes of various human cardiovascular diseases, with cardiovascular disease remaining a significant cause of mortality and disability in the U.S.
Melkani and his colleagues at UAB tested their model on aging fruit fly hearts and a fruit fly model with dilated cardiomyopathy, showing promising automated assessments validated against existing experimental data. Their model accurately tracked changes in cardiac parameters over time, from one week to five weeks old in fruit flies.
This model’s applicability extends to consumer hardware, providing calculated statistics such as diastolic and systolic intervals, fractional shortening, ejection fraction, heart rate, and quantified heartbeat irregularities.
Melkani highlights, “Our innovative platform for deep learning-based segmentation is the first of its kind for analyzing Drosophila hearts using standard high-resolution optical microscopy while quantifying all relevant parameters.”
By automating and enhancing the measurement process, this method enables more accurate and efficient studies on heart function in fruit flies, offering insights that could be applied to human cardiovascular research.
