SMU has developed SmartCADD, an open-source virtual tool that fuses artificial intelligence, quantum mechanics, and Computer Assisted Drug Design (CADD) approaches to accelerate the evaluation of chemical compounds, greatly shortening the time needed for drug discovery.
Discovering new drugs is akin to piecing together a jigsaw puzzle. The drug molecules must be carefully shaped to interact with the proteins in our bodies to create therapeutic effects. This intricate requirement makes the drug creation process exceptionally complex and lengthy.
To expedite this complex matching process, researchers at SMU have introduced SmartCADD, a virtual tool that merges artificial intelligence, quantum mechanics, and CADD methodologies to enhance the screening speed of chemical compounds, thus significantly decreasing the timeline for drug discovery. A recent study featured in the Journal of Chemical Information and Modeling showcased SmartCADD’s capability to identify potential drug candidates for HIV.
This innovative tool emerged from a collaborative effort between the chemistry department at SMU’s Dedman College of Humanities and Sciences and the computer science department at the Lyle School of Engineering.
“The need for new classes of medications, such as antibiotics, cancer therapies, and antivirals, is pressing,” noted Elfi Kraka, leader of the Computational And Theoretical Chemistry Group (CATCO) at SMU. “Even though AI is being rapidly integrated across many domains, there has been reluctance to utilize it in scientific inquiry, primarily due to concerns over data transparency and the quality of the data being employed for training. SmartCADD mitigates these issues and can analyze billions of chemical structures in just one day, significantly cutting down the time required to discover viable drug candidates.”
How SmartCADD Operates
SmartCADD employs deep learning models, systematic filtering protocols, and explainable AI to investigate extensive databases of chemical compounds aimed at identifying potential drug leads. It comprises two key elements: SmartCADD’s Pipeline Interface, which manages data collection and filtering, and its Filter Interface, which dictates the functionality of each filter. These embedded filters support different phases of chemical compound evaluation. They can forecast a drug’s behavior in the body, simulate drug structures using both 2D and 3D parameters, and utilize an AI model to clarify its decision-making process.
Researchers illustrated the capabilities of SmartCADD through three different case studies involving drugs for HIV treatment. They discovered several viral proteins that may serve as promising targets. SmartCADD harnessed data from the MoleculeNet library to construct and sift through a database comprising 800 million chemical compounds, concluding that 10 million had potential as HIV therapeutics. Subsequently, it utilized filters to pinpoint the compounds that best corresponded with established HIV medications.
Although the research concentrated on HIV in this study, the investigation emphasized that SmartCADD is adaptable and can be utilized in various drug discovery pipelines.
“This user-friendly virtual screening tool equips researchers with an integrated and flexible framework for developing drug discovery strategies,” expressed Corey Clark, an assistant professor of computer science at Lyle School of Engineering and deputy director for Research at SMU Guildhall. “We intend to continue advancing our work to further enhance chemistry and machine learning capabilities. The potential of this project is genuinely thrilling, and I am confident the next stage will represent an even more significant advancement.”
The Power of Collaboration for SmartCADD
The research paper also underscores the effectiveness of interdisciplinary collaboration at SMU. In addition to Kraka and Clark, contributors include Ayesh Madushanka, a postdoctoral fellow in chemistry, whose contributions are funded by a grant from the O’Donnell Data Science & Research Computing Institute, and computer science graduate student Eli Laird, a recipient of an O’Donnell Institute Ph.D. fellowship.
“Drug discovery necessitates teamwork for true success,” Madushanka remarked. “Had only the chemistry department worked on this project, the final outcome would likely have been different. Interdisciplinary collaboration introduces new perspectives on the same concept, refining and enhancing the result.”
Laird added, “Collaborative research across disciplines is essential for achieving significant advancements that have real-world impacts. This focus at SMU is a major reason behind my decision to pursue a Ph.D. here. Impactful research cannot thrive in isolation; it requires a broad, interdisciplinary view to inspire ideas that can evolve into genuine innovations. Many breakthroughs emerge where various fields intersect, and my goal is to position my research in such intersections.”
Funding for this research was provided by the National Science Foundation under grant CHE 2102461. The views, findings, and conclusions expressed in this material are those of the authors and do not necessarily reflect the National Science Foundation’s standpoint.
