Search for: All records

Cardiovascular diseases (CVD) are the leading cause of death worldwide. Engineered heart tissue produced by differentiation of human induced pluripotent stem cells may provide an encompassing treatment for heart failure due to CVD. However, considerable difficulties exist in producing the large number of cardiomyocytes needed for therapeutic purposes through differentiation protocols. Data-driven modeling with machine learning techniques has the potential to identify factors that significantly affect the outcomes of these differentiation experiments. Using data from previous cardiac differentiation experiments, we have developed data-driven modeling methods for determining which experimental conditions are most influential on the final cardiomyocyte content of a differentiation experiment. With those identified conditions, we were able to build classification models that can predict whether an experiment will have a sufficient cardiomyocyte content to continue with the experiment on the seventh (out of 10) day of the differentiation with a 90% accuracy. This early failure prediction will provide cost and time savings, as each day the differentiation continues requires significant resources. 

Surrogate models are used to map input data to output data when the actual relationship between the two is unknown or computationally expensive to evaluate. We have constructed a tool to recommend the appropriate surrogate modelling technique for a given dataset using attributes calculated from the input and output values. The tool identifies the appropriate surrogate modeling techniques with an accuracy of 98% and a precision of 91%. 

Surrogate models are used to map input data to output data when the actual relationship between the two is unknown or computationally expensive to evaluate for sensitivity analysis, uncertainty propagation and surrogate based optimization. This work evaluates the performance of eight surrogate modeling techniques for design space approximation and surrogate based optimization applications over a set of generated datasets with known characteristics. With this work, we aim to provide general rules for selecting an appropriate surrogate model form solely based on the characteristics of the data being modeled. The computational experiments revealed that, in general, multivariate adaptive regression spline models (MARS) and single hidden layer feed forward neural networks (ANN) yielded the most accurate predictions over the design space while Random Forest (RF) models most reliably identified the locations of the optimums when used for surrogate-based optimization. 

Abstract This study employed machine learning (ML) models to predict the cardiomyocyte (CM) content following differentiation of human induced pluripotent stem cells (hiPSCs) encapsulated in hydrogel microspheroids and to identify the main experimental variables affecting the CM yield. Understanding how to enhance CM generation using hiPSCs is critical in moving toward large‐scale production and implementing their use in developing therapeutic drugs and regenerative treatments. Cardiomyocyte production has entered a new era with improvements in the differentiation process. However, existing processes are not sufficiently robust for reliable CM manufacturing. Using ML techniques to correlate the initial, experimentally specified stem cell microenvironment's impact on cardiac differentiation could identify important process features. The initial tunable (controlled) input features for training ML models were extracted from 85 individual experiments. Subsets of the controlled input features were selected using feature selection and used for model construction. Random forests, Gaussian process, and support vector machines were employed as the ML models. The models were built to predict two classes of sufficient and insufficient for CM content on differentiation day 10. The best model predicted the sufficient class with an accuracy of 75% and a precision of 71%. The identified key features including post‐freeze passage number, media type, PF fibrinogen concentration, CHIR/S/V, axial ratio, and cell concentration provided insight into the significant experimental conditions. This study showed that we can extract information from the experiments and build predictive models that could enhance the cell production process by using ML techniques. 

Abstract Cardiovascular disease is the leading cause of death worldwide, and current treatments are ineffective or unavailable to majority of patients. Engineered cardiac tissue (ECT) is a promising treatment to restore function to the damaged myocardium; however, for these treatments to become a reality, tissue fabrication must be amenable to scalable production and be used in suspension culture. Here, we have developed a low‐cost and scalable emulsion‐based method for producing ECT microspheres from poly(ethylene glycol) (PEG)–fibrinogen encapsulated mouse embryonic stem cells (mESCs). Cell‐laden microspheres were formed via water‐in‐oil emulsification; encapsulation occurred by suspending the cells in hydrogel precursor solution at cell densities from 5 to 60 million cells/ml, adding to mineral oil and vortexing. Microsphere diameters ranged from 30 to 570 μm; size variability was decreased by the addition of 2% poly(ethylene glycol) diacrylate. Initial cell encapsulation density impacted the ability for mESCs to grow and differentiate, with the greatest success occurring at higher cell densities. Microspheres differentiated into dense spheroidal ECTs with spontaneous contractions occurring as early as Day 10 of cardiac differentiation; furthermore, these ECT microspheres exhibited appropriate temporal changes in gene expression and response to pharmacological stimuli. These results demonstrate the ability to use an emulsion approach to encapsulate pluripotent stem cells for use in microsphere‐based cardiac differentiation.