An Open-Source Database for Predicting Pharmacokinetics (360G-Wellcome-214464_Z_18_Z)

A free public web resource will be created for academic and industry pharmacometric modellers and data scientists to share standard big datasets and code for the purpose of generating reproducible studies. The ultimate goal is to encourage open research practices through provision of resources needed to facilitate machine learning and mechanism-based approaches for modelling of dose-concentration-response. The repository will allow data users to interact with data generators promoting research collaboration and sharing with a wider audience. The completed project will be a website at the address: www.pkpdai.com and will have 3 overarching aims: Data access: Links to PKPD datasets amenable to machine learning, the first being clearance (CL) or CL/F for non intravenous administration, where F is the bioavailability. Fostering collaboration: To download the data and use analytics users will need to register. Upon logging in users will see a simple profile of other users: Name; Institution; Country with an option to send a message to that user (modelled on the PAGE website participant list). Project Showcase: Summaries, links and other material relating to projects using the database. Milestones: 1. Month 6: Website functional with pubmed updating system established, paper submitted to Wellcome Open Research 2. Month 7: Workshop at PAGE meeting, email dissemination and set up a Kaggle competition for machine learning prediction of CL. 3. Month 8: downloads by > 3 academic groups and > 3 pharma companies 4. Month 12 (end of grant): Workshop at ASCPT, > 50 registered users, 1 manuscript submitted on using database for CL prediction.