# Designing Personalized Formulations with Machine Learning

> **NIH NIH R35** · DUKE UNIVERSITY · 2024 · $322,000

## Abstract

ABSTRACT
The design of drug formulations is an essential part of pharmaceutical development to enable the safe and
effective delivery of medications. Unfortunately, formulation optimization is currently done using a trial-and-error
approach or by adhering to already established formulation strategies following a one-size-fits-all mindset. This
has resulted in formulations that are simple and only ensure appropriate physical properties such as shelf life
and liberation. Complex, targeted formulations can increase the safety and efficacy of medications, but such
systems are expensive to design, manufacture, and administer – limiting their broader deployment. Here, we
describe our goals to expand and augment our efforts in developing innovative machine learning methods and
integrate them with experimental workflows for the design of novel, targeted drug formulations. We will
specifically focus on the machine learning-guided design of (1) functional excipients that prevent microbiome
metabolism, (2) targeted self-assembling nanoparticles, and (3) tissue-selective prodrugs. Our machine learning
models will enable us to circumvent billions of otherwise necessary trial-and-error experiments by predicting the
most promising candidates for experimental validation. This allows us to systematically explore novel drug
delivery systems to identify better solutions that work best for specific medications and patients. Our focus on
functional excipients, self-assembled nanoparticles, and prodrugs will provide delivery solutions that are easier
to produce and deploy on a larger scale, thereby enhancing the impact of advanced drug delivery systems and
making medicine more equitable. Our in vitro and in vivo experiments will validate our predictions and provide
pre-clinical data for innovative drug delivery solutions positioned for further translation. We expect that our
platform will enable the rapid and effective design of advanced drug delivery solutions to create safer and more
efficacious therapeutics for every patient.

## Key facts

- **NIH application ID:** 10890751
- **Project number:** 5R35GM151255-02
- **Recipient organization:** DUKE UNIVERSITY
- **Principal Investigator:** Daniel Reker
- **Activity code:** R35 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2024
- **Award amount:** $322,000
- **Award type:** 5
- **Project period:** 2023-08-01 → 2028-05-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10890751

## Citation

> US National Institutes of Health, RePORTER application 10890751, Designing Personalized Formulations with Machine Learning (5R35GM151255-02). Retrieved via AI Analytics 2026-05-23 from https://api.ai-analytics.org/grant/nih/10890751. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*