AUTHOR=Jamal Salma , Grover Abhinav , Grover Sonam TITLE=Machine Learning From Molecular Dynamics Trajectories to Predict Caspase-8 Inhibitors Against Alzheimer’s Disease JOURNAL=Frontiers in Pharmacology VOLUME=10 YEAR=2019 URL=https://www.frontiersin.org/journals/pharmacology/articles/10.3389/fphar.2019.00780 DOI=10.3389/fphar.2019.00780 ISSN=1663-9812 ABSTRACT=

Alzheimer’s disease (AD) is a neurodegenerative disorder in which the death of brain cells takes place leading to loss of memory and decreased cognitive ability. AD is a leading cause of death worldwide and is progressive in nature with symptoms worsening over time. Machine learning–based computational predictive models based on 2D and 3D descriptors have been effective in identifying potential active compounds. However, the use of data from molecular dynamics (MD) trajectories for training machine learning models still needs to be explored. In the present study, descriptors have been extracted from the MD trajectories of caspase-8 ligand complexes to train models using artificial neural networks and random forest algorithms. Caspase-8 plays a key role in causing AD by cleaving amyloid precursor proteins during apoptosis leading to increased formation of the amyloid-beta peptide. A total of 43 ligands were docked using the glide module of Schrodinger software, and short MD simulations of 10 ns were performed for the calculation of MD descriptors. The MD descriptors were also combined with the 2D and 3D descriptors of chemical compounds, and individual descriptor based as well as combination models were generated. This study demonstrated that MD descriptors could be effectively used for the characterization of bioactive compounds along with lead prioritization and optimization.