Drug Delivery System Tailoring Via Metal-Organic Framework Property Prediction Using Machine Learning: A Disregarded Approach Publisher



Pouyanfar N¹ ; Ahmadi M² ; Ayyoubzadeh SM^{3, 4} ; Ghorbanibidkorpeh F¹ Show Affiliations
Source: Materials Today Communications Published:2024

Abstract

Metal-organic frameworks (MOFs) have demonstrated exclusive features, including high porosity, high surface area, favorable biodegradability, and biocompatibility. Due to these unique properties, MOFs could be used extensively in several applications, including gas storage, catalyst, separation, and biomedical applications. Also, multiple types of MOFs with various features have been identified and used in many applications. Due to the diversity in the characteristics and capabilities of MOFs, their experimental examination is complex, costly, and time-consuming. Machine learning (ML), as an artificial intelligence tool, is an alternative for accurately predicting MOF characteristics. MOFs can benefit from integrating ML techniques for the design and development of them in various fields. This review summarizes ML tools that can be applied to predict MOF properties in designing drug delivery systems (DDSs). To this end, various ML classifiers were introduced first, and then related studies of ML methods for predicting MOF properties and capabilities were presented. ML exhibited a unique role in MOF research to predict the properties of MOFs with less cost and time. Finally, the potential application of ML in developing optimum MOF-based drug carriers is presented as a route to in silico DDS tailoring. © 2024 Elsevier Ltd