An open-source database for predicting pharmacokinetics

Project Summary 

Our aim is to create a web resource for academic and industry modellers in the field of pharmacology to share standard big datasets and methodological techniques. The ultimate goal with the output is to encourage open research practices in clinical pharmacology through providing the data resources needed to make a shift from a reductionist (semi) mechanism-based approach to a machine learning based approach for modelling of dose-concentration-response.

The completed project will be a website at the address: www.pkpdai.com. The first dataset will focus on the pharmacokinetic parameters clearance (CL) or CL/F, for non-intravenous administration, where F is the bioavailability.  We are developing a system to automatically expand this dataset as new papers are published.  A data dictionary will be provided and simple data querying and plotting tools made available via R Shiny. The website will then have the capacity to grow by adding new datasets and model code.

Key Project Information

Dates: 01/08/2019 - 31/07/2020

Status: Active

UCL lead/Principal Investigator: Joseph Standing, Infection, Immunity & Inflammation Dept, UCL GOS Institute of Child Health

IGH lead: Frank Kloprogge

Partner: BenevolentAI

Location: UK

Funding: Wellcome Trust

Contact: j.standing@ucl.ac.uk

Website: www.pkpdai.com/

Publications

Gonzalez Hernandez F, Carter SJ, Iso-Sipilä J et al. An automated approach to identify scientific publications reporting pharmacokinetic parameters [version 1; peer review: 1 approved]. Wellcome Open Res 2021, 6:88 (https://doi.org/10.12688/wellcomeopenres.16718.1)