Back

Predicting Molecular Docking Affinity of Per- and Polyfluoroalkyl Substances (PFAs) Towards Human Blood Proteins Using Generative AI Algorithm DiffDock

Fortela, D. L. B.; Mikolajczyk, A. P.; Carnes, M.; Sharp, W.; Revellame, E.; Hernandez, R.; Holmes, W.; Zappi, M.

2023-08-06 bioinformatics
10.1101/2023.08.03.551898 bioRxiv
Show abstract

This study computationally evaluates the molecular docking affinity of various perfluoroalkyl and polyfluoroalkyl substances (PFAs) using a generative machine learning algorithm, DiffDock, specialized in protein-ligand blind-docking learning and prediction. Concerns about the chemical pathways and accumulation of PFAs in the environment and eventually in human body has been rising due to empirical findings that levels of PFAs in human blood has been rising. Though there is currently a heightened need to understand the pathways of PFAs, empirical studies on PFAs have been relatively slow due to the time-scale and cost of standard chemical analysis such as those in blood samples. The current study demonstrates the implementation of DiffDock and assesses the prediction results in relation to empirical findings. The capability of an advanced generative artificial intelligence (AI) algorithm designed for protein-ligand docking such as DiffDock offers a fast approach in determining the potential molecular pathways of PFAs in human body.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.