高斯过程建模用于预测聚合物纳米粒子配方设计,以提高封装效率和治疗疗效
Sihan Dong1, Haolin Yu2, Pascal Poupart2
1School of Pharmacy, University of Waterloo, Ontario, N2G 1C5, Canada.
Drug delivery and translational research
|May 20, 2024
概括
机器学习,特别是高斯过程模型,可以加速纳米粒子 (NP) 药物递送系统的优化. 这些模型准确地预测了药物封装效率和治疗疗效对于多 (乳糖-同-甘油酸) (PLGA) NPs.
科学领域:
- 生物材料科学 生物材料科学
- 纳米技术纳米技术
- 计算化学的计算化学
背景情况:
- 传统的药物输送面临着诸如第一通代谢和非目标效应等挑战.
- 纳米粒子 (NP) 通过改善药物输送和减少副作用提供了一个解决方案.
- 优化NP材料成分至关重要,但往往耗时且昂贵.
研究的目的:
- 探索机器学习 (ML) 的使用,以加速纳米粒子 (NP) 组成优化.
- 开发高斯过程 (GP) 模型来预测药物封装效率 (EE%) 和NP的治疗疗效.
- 为了验证GP模型的预测准确性,使用聚 (乳酸-co-糖醇酸) (PLGA) 纳米粒子配方.
主要方法:
- 制造了32种具有不同物理化学性质的聚 (乳糖合甘油酸) (PLGA) 纳米粒子配方.
- 装载着模型药物的NP:多克索鲁比辛 (DOX) 和多塞塔克塞尔 (DTX).
- 利用高斯过程 (GP) 模型预测EE%和IC50值 (治疗疗效) 对OVCAR3卵巢癌细胞.
主要成果:
- EE% GP模型实现了最高的预测准确性,正常化的RMSE为0.187.
- 药物疗效的GP模型显示DOX的0.296和DTX的0.206的正常化RMSE.
- 这些结果表明了GP模型在预测NP性能方面的潜力.
结论:
- 高斯过程模型可以显著加速纳米粒子药物输送系统的优化.
- 基于ML的EE%和治疗疗效的预测是可行的和准确的.
- 这种方法可以减少与开发新型NP药物输送配方相关的时间和成本.
相关概念视频
Factors Affecting Dissolution: Particle Size and Effective Surface Area
807
Dissolution kinetics, an essential aspect of oral drug delivery, is significantly influenced by the drug's particle size. According to the Noyes-Whitney dissolution model, the dissolution rate correlates directly with the drug's surface area. The larger the surface area, the higher the drug's solubility in water, leading to a faster drug dissolution rate. Reducing particle size increases the effective surface area, enhancing the dissolution process. Micronization and nanosizing are...
807
Factors Influencing Drug Absorption: Pharmaceutical Parameters
129
Solid dosage forms such as tablets and capsules undergo rigorous manufacturing processes to ensure stability and effectiveness. Their dissolution and absorption properties are influenced significantly by the choice of excipients (inactive ingredients that serve various roles in the formulation), and the methodology applied during production. The manufacturing parameters, such as compression force and granulation techniques, significantly affect dissolution rates. Elevated compression forces...
129


