Artificial Intelligence in Drug Discovery
1
2021-2022
02042311
Not specified
English
B-learning
1.0
Elective
Non Degree Course
Recommended Prerequisites
Not applicable.
Teaching Methods
Multiple-answer/choice questions.
Learning Outcomes
After the lecture (and studying of the provided materials), the student should be able to:
– list 5-7 challenges facing the Drug Discovery field where Artificial Intelligence may help overcoming
– describe in reasonable depth a few (2-3) examples wherein an AI methodology has delivered improvements over the state-of-the-art in silico techniques
– demonstrate a reasonable understanding as to why certain AI methodologies are more adequate (than others) in solving specific Drug Discovery problems
– provide examples of web resources and online data repositories that are most relevant to the drug discovery field
– (OPTIONAL): describe the process of training and testing a molecular property predictor using Google CoLab.
Work Placement(s)
NoSyllabus
– [INTRODUCTION] The Drug Discovery Process and Artificial Intelligence (AI): Moore meets Eroom
– Pharmacological target fishing, off-target interaction predictions and polypharmacology: the harnessing of Big (Bioactivity) Data
– Virtual Screening today: the Computational Chemistry–Cheminformatics–Machine Learning (ML) “love triangle”
– Molecular property predictions from “small” data using “simple” ML: Solubility and Permeation as case studies
– (OPTIONAL) Hands-On Session: modeling aqueous solubility using RDKit and a IPython Notebook in Google CoLab
– Generating novel molecules with desired properties using Deep (Reinforcement) Learning
– AI-informed COVID-19 drug repositioning: perspectives, challenges and directions
– AI for patient stratification and personalized therapies: are we there yet?
– [ENDING REMARKS] Ten rules to boost drug discovery with data science: the Novartis perspective.
Head Lecturer(s)
Jorge António Ribeiro Salvador
Assessment Methods
Assessment
Multiple-answer/choice questions: 100.0%
Bibliography
A brief collection of keyword lookups before the lecture:
– PubChem, ChEMBL, RCSB Protein Data Bank, ZINC database, DrugBank
– Innovative Medicines Initiative (IMI), The Melloddy project, Drug Discovery Catapult
– Click2Drug, RDKit, Conda, IPython Notebooks, Google CoLab, R (CRAN)
– Benevolent.AI, ExScientia, AtomWise, InSilico Medicine
References:
– DOI: https://doi.org/10.1038/s41563-019-0338-z
– DOI: https://doi.org/10.1093/bib/bby061