Using DNA-Encoded Chemical Libraries to Enable Machine Learning for Alzheimer's Disease Drug Discovery

This project will generate tool compounds and chemical probes for putative therapeutic targets in Alzheimer’s Disease (AD) and AD-related dementias (ADRD) using DNA encoded libraries (DELs) and machine learning (ML). DELs are combinatorial small molecule libraries in which the identity of each small molecule is encoded in a unique, covalently attached DNA tag. We will use our pre-existing DELs at the Baylor College of Medicine Center for Drug Discovery to identify hits against dozens of AD/ADRD targets and to produce data sets with which we will train DEL-ML models to predict binders to these targets. We will share the compounds we make and a subset of the selection data to encourage ML method development and AD drug discovery by others. In Aim 1, we will perform DEL selections on putative AD/ADRD targets selected in consultation with our Target Advisory Board. We will share the selection input of 12 DELs (907 million structures: our OpenDEL) and the output (sequencing counts for tens of thousands of hits per target) as public data sets. We will synthesize high- ly enriched DEL hits off DNA, test them for activity, and share the data publicly. We will use focused DELs, as described in Aim 3, to optimize biochemically validated scaffolds and share these tool compounds. We will train DEL-MLs to identify hits in virtual re-purposing libraries composed of investigational small-molecule drugs that have passed a Phase I clinical trial, purchase hits, and test them biochemically. True positives in this drug re-purposing could be tested in animals and immediately go into clinical trials. In Aim 2, we will pursue in greater depth a few targets chosen with our Target Advisory Board from our Aim 1 portfolio. We will synthesize hits, test these for biochemical activity or binding, and pursue hit to lead optimi- zation. If our hits show toxicity or fail to cross the blood-brain barrier, we will use our DEL-MLs to identify unrelated hits in a compound virtual library curated for drug-like molecular properties and brain penetrance. We will purchase dozens of these compounds and test them for binding or biochemical activity. We will increase the potency of validated novel binders by constructing new DELs that incorporate the novel scaffolds. We will optimize the molecular properties of potent, selective hits, including the ability to pass the BBB, and test the efficacy of promising leads in animal models via collaboration. Aim 3 outlines how we will design and synthesize focused DELs or novel DELs that explore chemical space around a novel scaffold. We will use the predictive power of our DEL-MLs to assess our designs before implementing them synthetically. Selections performed using these DELs will provide potent hits to be tested in Aims 1 and 2, and large amounts of data to better train our DEL-MLs. Project Number: 1R01AG098158-01 | Fiscal Year: 2026 | NIH Institute/Center: National Institute on Aging (NIA) | Principal Investigator: Damian Young | Institution: BAYLOR COLLEGE OF MEDICINE, HOUSTON, TX | Award Amount: $1,343,997 | Activity Code: R01 | Study Section: Special Emphasis Panel[ZRG1 NV-L (52)] View on NIH RePORTER: https://reporter.nih.gov/project-details/11294057