Machine Learning Identifies Abnormal Ca2+ Transients in Human Induced Pluripotent Stem Cell-Derived Cardiomyocytes

­ Application NAC (N-acetyl-L-cystine) as an antioxidant to reduce chemotherapy induced oxidative stress and cardiotoxicity. Key Benefits Potentially decreases the time and labor for evaluating transient calcium compared to current gold stand…

Application
NAC (N-acetyl-L-cystine) is an antioxidant to reduce chemotherapy-induced oxidative stress and cardiotoxicity.

Key Benefits

  • Potentially decrease the time and labor for evaluating transient calcium compared to current gold standard methods.
  • May accelerate the development of new drugs for cardiovascular diseases.
  • Could improve the basic understanding of cardiovascular diseases as an essential research tool.


Market Summary
Heart disease is the leading cause of death in the U.S., killing one person every 36 seconds (CDC). Abnormal transient calcium is a known risk factor for multiple heart conditions, but manual identification and quantification of abnormal calcium is inefficient. The herein described machine learning approach using human pluripotent stem cell training data to evaluate and measure transient calcium can mitigate bottlenecks and enable automatic assessment of Ca2+ transient abnormality. This invention would contribute to the global cardiovascular disease diagnostics & monitoring market, which is expected to reach $10.14 billion by 2022 at a CAGR of 4.0% from 2016-to 2022 (BCC Research HLC210A).

Technical Summary
Researchers have developed a novel machine-learning approach to evaluate and measure transient calcium utilizing human pluripotent stem cell training data. The current manual identification pf abnormal Ca(2+) transients is labor-intensive and subjective to the assessor’s expertise. To address this the researchers adapted an extant machine learning-based Ca(2+) transient peak analysis algorithm and vastly improved its capability and accuracy in Ca(2+) transient peak identification and variables quantification. The researchers found using two separate Support Vector Machine classifiers that when trained against human accessors the novel computational tool demonstrated an improved level of accuracy, sensitivity, and specificity when assessing peak abnormality and cell abnormality.

Developmental Stage
Early-stage. Publication: Hwang, H., Liu, R., Maxwell, J.T. et al. Machine learning identifies abnormal Ca 2+ transients in human induced pluripotent stem cell-derived cardiomyocytes. Sci Rep 10, 16977 (2020). https://doi.org/10.1038/s41598-020-73801-x (October 12, 2020)