In the early stage of small molecule drug discovery, accurate prediction of the binding behavior between the target protein and the compounds is crucial for discovering the candidate molecule with decent potency and selectivity. In practical situations, drug-target binding affinity (DTA) prediction becomes difficult because of the lack of a precise protein structure model or known ligands. Data-driven methods are actively being used to conquer the problem. However, the available data is usually insufficient to build a model with generalization capability for the novel targets.
A promising approach to remedy these situations is a few-shot prediction of drug-target bindings, which relies on the small number of the known binding affinities of compounds. A neural process-based model for this purpose (MetaDTA) was recently developed, which does not use the structural information of the proteins. The method is based on meta-learning the known binding affinity values for various protein targets. The prediction accuracy is superior to the state-of-the-art structure-free prediction models, and the model is highly generalizable for unknown targets.
Join this webinar to learn about the proposed meta-learning approach for drug-target binding affinity prediction with the results, limitations and future perspectives.
Jiho Yoo, AI Director, Standigm Inc.
As an AI Director of Standigm, Jiho Yoo, PhD, is focusing on the development of Standigm’s AI module technologies. Prior to Standigm, Jiho developed AI algorithms related to organic materials and automatic synthesis at Samsung Advanced Institute of Technology (SAIT). He received a PhD in computer science (machine learning) from Pohang University of Science and Technology (POSTECH). He is a co-author of several publications about molecular design and machine learning.
Who Should Attend?
This webinar will appeal to researchers, scientists, academicians and professionals from pharmaceutical companies, biopharma companies as well as universities and research institutes with interests in adopting artificial intelligence (AI) in target research for diverse indications.
Professional roles or responsibilities including, but not limited to:
- Chief Scientific Officers
- Project Managers
- Principal Investigators
- Research Scientists
- Research Directors
- Post-Doctoral Researchers
- Research Associates
- Directors of Research Development
What You Will Learn
In the early stage of small molecule drug discovery, the data is always limited to accurately model the required properties. Meta-learning can be a promising approach for this situation, especially in the drug-target binding affinity predictions.
Standigm is a workflow AI-driven drug discovery company headquartered in Seoul, South Korea and subsidized in US and UK. Standigm has proprietary AI platforms encompassing novel target identification to compound design, to generate commercially valuable drug pipelines. Founded in 2015, Standigm has established an early stage drug discovery workflow AI to generate multiple First-in-Class compounds within seven months. Pursuing full-stack, AI-driven industrializing drug discovery, Standigm has achieved the automation of molecular design workflow, and the automation effort has been expanding to the whole drug discovery process based on Standigm AI platforms, including Standigm ASK(tm) for novel target discovery, Standigm BEST(tm) for novel compound generation.