Our group at ETH Zürich has recently produced novel dual-display DELs with diverse encoding schemes and innovative chemical designs, including fragment-like small molecules and several macrocyclic architectures. Such libraries can be mixed-and-matched together to reach a higher level of combinatorial assembly. Potent binders were successfully obtained for a large array of targets, demonstrating the yet untapped potential of dual-display DELs for ligand discovery against important therapeutic targets.

Over the last several years, DNA-encoded library (DEL) screens have become a critical part of Johnson & Johnson’s integrated hit-finding workflow. We will highlight recent advancements and applications of our internal DEL platform including several recently developed DEL-compatible chemistries that enable expansion of DEL-accessible chemical space, key DEL platform metrics, and successful hit-ID campaigns featuring DEL screens.

This presentation will explore the DEL selection campaign we designed to screen our internal library collection of ~5 billion molecules to discover novel and potent PARP1 inhibitors that do not exhibit toxic DNA trapping properties. The selection and protein design allowed us to interrogate an active, functionally relevant form of PARP1; the development of biochemical assays, alongside obtaining a crystal structure, enabled the validation of the MoA of these inhibitors.

We used a DNA-encoded chemistry technology (DEC-Tec) to discover inhibitors of SARS-CoV-2 main protease (Mpro) as an alternative to current strategies. The development of inhibitors for the treatment of COVID-19 has mostly benefitted from X-ray structures and preexisting knowledge of inhibitors. I will discuss how our approach provides an effficient method to generate Mpro inhibitors by circumventing such limitations.

Transcription factors and other sequence specific DNA binding proteins pose unique challenges for DNA Encoded library screening, where DNA-driven binding of DEL may lead to false positive hits. Nurix has combined multiple strategies to address this challenge and identify hits against the DNA-binding domain of the EWS-FLI1 fusion oncoprotein. This approach is widely applicable to DEL screening of DNA binding proteins.

DEL screens typically entail affinity selection to discover ligands of the protein target. Our laboratory has previously shown that solid-phase DELs can be screened directly for biochemical activity in microfluidic droplets. In this talk, we discuss polymer engineering and 3D culture principles that have enabled DEL screening on the basis of cellular activity. 

This talk will describe microfluidics-enabled cellular phenotypic DEL workflow—MicDrop. We will first introduce molecular glue degrader concept with an inspirational example, and describe its potential implications and how phenotypic DEL platform could unlock new opportunities. We will then share how we overcame the challenges to perform cellular DEL screen in droplets, followed by results from a cellular protein degradation screen with a validation library. Our results show the benefits of bead replicates and how this new paradigm of DEL screen can accelerate the field of molecular glue discovery for TPD and beyond.

We present novel approaches for the selection of molecules from DNA-encoded libraries using enzymatic tags on target proteins. We apply these assays for DEL discovery with GPCRs in live cells for both the discovery of ligands and for specific discovery of biased agonists.

Treatments for PIK3CA-mutant cancers are limited by toxicities associated with the inhibition of WT PI3Ka. I will describe how Relay Therapeutics integrates our Dynamo Platform with DEL screening to identify mutant-selective chemical starting points. From these starting points came the development of RLY-2608, a first-in-class inhibitor demonstrating mutant selectivity in patients.

DNA-Encoded Library (DEL) technology provides a high-throughput and cost-effective screening approach for lead discovery. While the strategies for DEL screening and data analysis have greatly improved, data normalization remains an open challenge. Existing normalization methods can yield poor correlation for compounds with high copy-count, and they do not account for inherent sources of noise. To overcome these drawbacks, we have developed a robust normalization technique using an anti-HA antibody and on-DNA HA peptide as a positive control pair. This new approach allows for normalization between samples of different conditions and accounts for technical challenges that occur during screening.

I will provide a critical review of examples in the literature and industry to determine if, how, when and why using AI or machine learning has helped drug lead generation when using DNA-Encoded Libraries.

At Anagenex, we are coupling the high-throughput power of DNA-encoded libraries with Machine Learning in order to drive the hit-to-lead process. We will discuss a case study wherein we were able to drive a target campaign via leveraging focused libraries with Machine Learning to enable rapid chemotype expansion and decision making.

Machine learning models make better predictions of small molecule binders to proteins when they are built on better training sets. Training sets enable better models when they (i) comprise more, and diverse, true positives and negatives, and (ii) when the true positives are more accurately rank-ordered by affinity. We are building DELs and DEL selection methods that produce higher-quality training sets.