Wystąpienie w ramach Poznańskiego Horyzontu Danych:
—
Opis:
In this short presentation, I will show how we can apply machine learning in drug discovery for molecule property prediction. As an example use case, we will try to predict the hERG-related cardiotoxicity of potential drug candidates. We will use Autogluon (AutoML framework) for tabular data to build and select the optimal model. We will go through each step, starting from molecule data preparation for machine learning tasks through data exploration, model training, and model selection. I will comment on the advantages and disadvantages of autoML frameworks. I hope you will be able to use the techniques presented and apply them beyond drug discovery and the chemoinformatics field.
Bio:
Marcin Kowiel is an accomplished professional with a diverse background in computer science, mathematics, and machine learning. He earned his PhD in small molecule crystallography from Poznan University of Medical Sciences in 2015 and continued research in protein crystallography at the Institute of Bioorganic Chemistry, Polish Academy of Sciences. Proficient in Python development and data science, since 2018 he has led teams in cyber security company (F-Secure) and genomics-focused startup (MNM Diagnostics). Currently as Data Science Team Leader at Ryvu Therapeutics, Marcin leverages AI for drug discovery, with a primary focus on hit identification and lead optimization stages.
source
