Welcome to the new version of CaltechAUTHORS. Login is currently restricted to library staff. If you notice any issues, please email coda@library.caltech.edu
Published December 7, 2020 | Submitted
Report Open

Accelerated Predictive Healthcare Analytics with Pumas, A High Performance Pharmaceutical Modeling and Simulation Platform

Abstract

Pharmacometric modeling establishes causal quantitative relationships between administered dose, tissue exposures, desired and undesired effects and patient's risk factors. These models are employed to de-risk drug development and guide precision medicine decisions. However, pharmacometric tools have not been designed to handle today's heterogeneous big data and complex models. We set out to design a platform that facilitates domain-specific modeling and its integration with modern analytics to foster innovation and readiness in healthcare. Pumas demonstrates estimation methodologies with dramatic performance advances. New ODE solver algorithms, such as coeficient-optimized higher order integrators and new automatic stiffness detecting algorithms which are robust to frequent discontinuities, give rise to a median 4x performance improvement across a wide range of stiff and non-stiff systems seen in pharmacometric applications. These methods combine with JIT compiler techniques, such as statically-sized optimizations and discrete sensitivity analysis via forward-mode automatic differentiation, to further enhance the accuracy and performance of the solving and parameter estimation process. We demonstrate that when all of these techniques are combined with a validated clinical trial dosing mechanism and non-compartmental analysis (NCA) suite, real applications like NLME fitting see a median 81x acceleration while retaining the same accuracy. Meanwhile in areas with less prior software optimization, like optimal experimental design, we see orders of magnitude performance enhancements over competitors. Further, Pumas combines these technical advances with several workflows that are automated and designed to boost productivity of the day-to-day user activity. Together we show a fast pharmacometric modeling framework for next-generation precision analytics.

Additional Information

The copyright holder for this preprint is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. Version 1 - November 30, 2020; Version 2 - March 20, 2022. The Pumas project was started in 2017 with the support of the University of Maryland Baltimore School of Pharmacy (UMB). Professor Jill Morgan, the chair of Pharmacy Practice and Science (PPS) and Dean Natalie Eddington were instrumental in their support. The staff, faculty, students and researchers at the Center for Translational Medicine in UMB drove a lot of the initial testing and we are extremely thankful for their patience while adopting a new language and tool. Brian Corrigan from Pfizer who encouraged us throughout the journey and nudged the team to achieve new heights. All the early adopters and believers. Viral Shah and Deepak Vinchii from Julia Computing for being our trusted technology collaborators. Last but the most important thank you goes out to the members of the Julia Language community who develop world class scientific computing packages and create an environment that is welcoming to newcomers. Competing Interest Statement: Pumas is a proprietary software developed by Pumas-AI Inc. Authors Rackauckas, Ma, Noack, Dixit, Mogensen, Elrod, Tarek, Byrne, Maddhashiya, Hatherly, Calderon, Nyberg, Gobburu, and Ivaturi all are or have been affiliated with Pumas-AI Inc. in the past 36 months.

Attached Files

Submitted - 2020.11.28.402297v2.full.pdf

Files

2020.11.28.402297v2.full.pdf
Files (3.4 MB)
Name Size Download all
md5:bcd803a37ee0acd86471f0f24ebc17b9
3.4 MB Preview Download

Additional details

Created:
August 20, 2023
Modified:
October 23, 2023