Department of Pharmaceutics, Faculty of Pharmacy, King Abdulaziz University (KAU), Jeddah, 21589, Saudi Arabia.
Blacktrace Holdings Ltd (Dolomite Microfluidics), Royston, SG8 5TW, UK.
Sci Rep. 2020 Nov 11;10(1):19517. doi: 10.1038/s41598-020-76477-5.
In this study, synthetic polymeric particles were effectively fabricated by combining modern technologies of artificial intelligence (AI) and microfluidics. Because size uniformity is a key factor that significantly influences the stability of polymeric particles, therefore, this work aimed to establish a new AI application using machine learning technology for prediction of the size of poly(D,L-lactide-co-glycolide) (PLGA) microparticles produced by diverse microfluidic systems either in the form of single or multiple particles. Experimentally, the most effective factors for tuning droplet/particle sizes are PLGA concentrations and the flow rates of dispersed and aqueous phases in microfluidics. These factors were utilized to develop five different and simple in structure artificial neural network (ANN) models that are capable of predicting PLGA particle sizes produced by different microfluidic systems either individually or jointly merged. The systematic development of ANN models allowed ultimate construction of a single in silico model which consists of data for three different microfluidic systems. This ANN model eventually allowed rapid prediction of particle sizes produced using various microfluidic systems. This AI application offers a new platform for further rapid and economical exploration of polymer particles production in defined sizes for various applications including biomimetic studies, biomedicine, and pharmaceutics.
在这项研究中，通过结合人工智能 (AI) 和微流控技术的现代技术，有效地制备了合成聚合物颗粒。由于粒径均匀性是显著影响聚合物颗粒稳定性的关键因素，因此，本工作旨在建立一种新的人工智能应用，使用机器学习技术来预测由不同微流控系统产生的聚(D,L-丙交酯-共-乙交酯) (PLGA) 微球的粒径，无论是单个颗粒还是多个颗粒的形式。实验中，调节液滴/颗粒尺寸的最有效因素是 PLGA 浓度以及微流控中分散相和水相的流速。这些因素被用于开发五个不同的、结构简单的人工神经网络 (ANN) 模型，这些模型能够分别或联合预测不同微流控系统产生的 PLGA 颗粒的粒径。ANN 模型的系统开发最终允许构建一个单一的计算机模型，该模型包含三个不同微流控系统的数据。该 ANN 模型最终允许快速预测使用各种微流控系统生产的颗粒粒径。这种人工智能应用为进一步快速和经济地探索用于各种应用的特定尺寸的聚合物颗粒生产提供了新的平台，包括仿生研究、生物医学和药剂学。
