Recent intellectual and engineering advances have helped the field progress from purely theoretical studies to the implementation of intelligent systems that solve problems in various aspects of our lives. As a multidisciplinary field, AI involves integrating insights from diverse disciplines such as computer science, mathematics, psychology, linguistics, philosophy, neuro-science, artificial psychology, and many others. (5) McCarthy’s original description still holds true today, albeit with some fleshing-out of the specifics. The term “artificial intelligence” was coined by John McCarthy at the Dartmouth Conference in 1956 to describe “the science and engineering of making intelligent machines”. Therefore, this review will focus on those aspects of this promising subfield that have already demonstrated their usefulness and applicability and reflect on those technologies which seem most promising for the next phase of AI in drug discovery. Thus, we will use the term “AI” as a synonym for certain machine learning techniques because there is no “strong” (general) AI to date. Machine learning approaches that might be considered instances of weak AI have made remarkable progress, with developments in both their fundamental algorithms and applications. In this context, machine learning and domain-specific (“weak”) artificial intelligence (AI) offer fresh opportunities for small-molecule drug discovery. One solution to such problems is to “outsource” our reasoning to a machine intelligence when it comes to the analysis of multisource and multidimensional data. Today, scientists have more information than ever before on a range of topics pertinent to the subject matter, far outpacing the ability of most to properly parse and integrate into their own workflows and research objectives. Here, we will focus on a central challenge in medicinal chemistry, namely the inherent difficulty of translating the process of drug discovery from basic science to early clinical trials. (3,4) The reasons for this tendency are manifold, including current market saturation, difficulties in bringing novel chemical matter through a complex approvals process, and the willingness-to-pay in developed and developing markets, among others. (1,2) The productivity of drug R&D (herein referring exclusively to small molecule drugs, unless otherwise specified) in the pharmaceutical industry remains on the decline, despite record expenditures. The research and development (R&D) cycle for innovative small-molecule drugs faces many challenges, such as high cost-to-market, limited success in clinical trials, and long cycle times. Finally, several challenges and limitations of the current methods are summarized, with a view to potential future directions for AI-assisted drug discovery and design. After introducing the basic principles, alongside some application notes, of the various machine learning algorithms, the current state-of-the art of AI-assisted pharmaceutical discovery is discussed, including applications in structure- and ligand-based virtual screening, de novo drug design, physicochemical and pharmacokinetic property prediction, drug repurposing, and related aspects. This review provides a comprehensive portrayal of these machine learning techniques and of their applications in medicinal chemistry. Various machine learning approaches have recently (re)emerged, some of which may be considered instances of domain-specific AI which have been successfully employed for drug discovery and design. Artificial intelligence (AI), and, in particular, deep learning as a subcategory of AI, provides opportunities for the discovery and development of innovative drugs.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |