Personal information

I am Van Thinh To, a ungraduate student in the Faculty of Pharmacy, University of Medicine and Pharmacy at Ho Chi Minh City. My research interests are applying machine learning and deep learning techniques to solving challenging problems in pharmaceutical chemistry, drug discovery, and computational chemistry. This is my blog, where I store my projects and some fundamental code for my research.

Research interests

My primary research interests involve:

• Development of Quantitative Structure-Activity Relationships (QSAR) models using machine learning algorithms
• Computer-aided drug design methodologies developement
• Molecular representation based on background model from unlabeled data

My research focuses on three main problems in drug discovery:

1. Molecular representation: how can we use unlabeled data (more than billions) to make features more generalizable? How can we vectorize steroisomerism information (conformation and ensemble conformation) from graph data?
2. Generative molecules: given a target (protein) and database of active-inactive molecules, how can we generate agonists/antagonists?
3. Retrosynthesis strategy: given a set of compounds and conditions, which compounds could be synthesized, and what is the probability of this reaction?

Our team - MedAI

Established in 2021 with students from the organic chemistry sector, we try to catch up with the innovation of artificial intelligence in drug discovery.