Bioscience Method 2024, Vol.15, No.1, 37-49 http://bioscipublisher.com/index.php/bm 39 of screening and validation, including repeated experiments under different experimental conditions and cross validation using different models. 1.2 The role and application of AI in drug screening AI (Artificial Intelligence) is playing an increasingly important role in drug screening. Its application not only improves screening efficiency, but also reduces research and development costs, bringing revolutionary changes to the field of drug development (Mohanty et al., 2020). AI can automate the processing and analysis of large-scale biomedical data, including genomics, proteomics, drug chemical structures, etc., in order to quickly identify potential drug targets related to specific diseases. This data mining and integration capability greatly surpasses traditional manual methods, providing researchers with more comprehensive and in-depth information. AI performs excellently in predicting drug target interactions. Hessler and Baringhaus (2018) explored the key role of AI in drug design, particularly how artificial neural networks such as deep neural networks drive various aspects of drug discovery, in their study. The application of AI in quantitative structure-activity relationship (QSAR) has demonstrated its strength in predicting physicochemical and ADMET (absorption, distribution, metabolism, excretion, and toxicology) properties. In addition, AI has demonstrated its strong ability to generate new bioactive molecules with desired properties, laying the foundation for its strength in drug discovery. Walters and Barzilay (2021) critically evaluated the application of AI in drug discovery in their review, exploring its application in analyzing high-content screening data, designing and synthesizing new molecules, and other drug discovery areas. They discussed the different fields in which AI is applied in drug discovery, including attribute prediction, molecular generation, image analysis, and organic synthesis planning. Deng et al. (2021) provided a review on the application and technology of AI in drug discovery. This study first provides an overview of drug discovery and its related applications, and then discusses common data resources, molecular representations, and benchmark platforms. AI technology is divided into model architectures and learning paradigms. They review the application of AI in drug discovery and provide a GitHub repository containing relevant papers (and applicable code) as a learning resource. Therefore, AI has the potential in drug screening and discovery, especially in molecular property prediction, molecular generation, and applications combined with synthesis planning and drug design. AI can also be used for virtual screening and experimental design, and even plays an important role in personalized treatment. 1.3 The importance of data mining and machine learning in drug development The importance of data mining and machine learning in drug development is self-evident. They provide a new perspective and tools for drug development, driving rapid development in the field of drug development. Data mining technology can extract information related to drug development from massive biomedical data. These data may come from multiple levels such as genomics, transcriptomics, proteomics, metabolomics, etc., covering rich content such as disease pathogenesis, drug action mechanisms, and clinical information of patients (Kavakiotis et al., 2017). Through data mining, researchers can gain a deeper understanding of the nature of diseases and the ways drugs work, providing strong data support for drug development. Machine learning algorithms can intelligently analyze and predict these data. Based on a large amount of training data, machine learning models can learn the complex relationship between drugs and diseases, predict key attributes such as the effectiveness and safety of candidate drugs. This predictive ability not only greatly improves the efficiency and accuracy of drug screening, but also helps to reduce the failure rate and risk in clinical trials. Data mining and machine learning can also provide personalized solutions for drug development. By conducting in-depth analysis of patient genomics, clinical data, etc., researchers can develop personalized treatment plans for each patient, improving treatment effectiveness and quality of life. The concept of precision medicine is gradually changing traditional medical models, bringing patients a better treatment experience.
RkJQdWJsaXNoZXIy MjQ4ODYzMg==