Machine-Learning-Based Data Analysis Method for Cell-Based Selection of DNA-Encoded Libraries

Enzymatic Cleavage of Double-Stranded DNA-Encoded Libraries (DELs) to Single-Stranded DELs with Compounds at the 3' End: Its Application in Photo-Crosslinking Selection

DNA-encoded library (DEL) technology is a crucial tool in pharmaceutical research, rapidly identifying compounds that bind to a target of interest from an extensive pool of compounds. In this study, we propose a new method for generating single-stranded DELs (ssDELs) with compounds at the 3' end. The introduction of uniquely designed hairpin-shaped headpieces containing deoxyuridine (NC-HP) and the use of a cleavage enzyme facilitate the conversion from double-stranded DELs (dsDELs) to such ssDELs. Moreover, Klenow fill-in provides the dsDELs with photo-crosslinkers covalently linked to the coding region, which exhibit durability even under stringent washing conditions and enable photo-crosslinking with a high signal-to-noise ratio, as also confirmed in cell-based photo-crosslinking selections.

Challenges and Prospects of DNA-Encoded Library Data Interpretation

DNA-encoded library (DEL) technology is a powerful platform for the efficient identification of novel chemical matter in the early drug discovery process enabled by parallel screening of vast libraries of encoded small molecules through affinity selection and deep sequencing. While DEL selections provide rich data sets for computational drug discovery, the underlying technical factors influencing DEL data remain incompletely understood. This review systematically examines the key parameters affecting the chemical information in DEL data and their impact on hit triaging and machine learning integration. The need for rigorous data handling and interpretation is emphasized, with standardized methods being critical for the success of DEL-based approaches. Major challenges include the relationship between sequence counts and binding affinities, frequent hitters, and the influence of factors such as inhomogeneous library composition, DNA damage, and linkers on binding modes. Experimental artifacts, such as those caused by protein immobilization and screening matrix effects, further complicate data interpretation. Recent advancements in using machine learning to denoise DEL data and predict drug candidates are highlighted. This review offers practical guidance on adopting best practices for integrating robust methodologies, comprehensive data analysis, and computational tools to improve the accuracy and efficacy of DEL-driven hit discovery.

Enhancing the Predictive Power of Machine Learning Models through a Chemical Space Complementary DEL Screening Strategy

DNA-encoded library (DEL) technology is an effective method for small molecule drug discovery, enabling high-throughput screening against target proteins. While DEL screening produces extensive data, it can reveal complex patterns not easily recognized by human analysis. Lead compounds from DEL screens often have higher molecular weights, posing challenges for drug development. This study refines traditional DELs by integrating alternative techniques like photocross-linking screening to enhance chemical diversity. Combining these methods improved predictive performance for small molecule identification models. Using this approach, we predicted active small molecules for BRD4 and p300, achieving hit rates of 26.7 and 35.7%. Notably, the identified compounds exhibit smaller molecular weights and better modification potential compared to traditional DEL molecules. This research demonstrates the synergy between DEL and AI technologies, enhancing drug discovery.

A Target Class Ligandability Evaluation of WD40 Repeat-Containing Proteins

Target class-focused drug discovery has a strong track record in pharmaceutical research, yet public domain data indicate that many members of protein families remain unliganded. Here we present a systematic approach to scale up the discovery and characterization of small molecule ligands for the WD40 repeat (WDR) protein family. We developed a comprehensive suite of protocols for protein production, crystallography, and biophysical, biochemical, and cellular assays. A pilot hit-finding campaign using DNA-encoded chemical library selection followed by machine learning (DEL-ML) to predict ligands from virtual libraries yielded first-in-class, drug-like ligands for 7 of the 16 WDR domains screened, thus demonstrating the broader ligandability of WDRs. This study establishes a template for evaluation of protein family wide ligandability and provides an extensive resource of WDR protein biochemical and chemical tools, knowledge, and protocols to discover potential therapeutics for this highly disease-relevant, but underexplored target class.

Agonist Discovery for Membrane Proteins on Live Cells by Using DNA-encoded Libraries

Identifying biologically active ligands for membrane proteins is an important task in chemical biology. We report an approach to directly identify small molecule agonists against membrane proteins by selecting DNA-encoded libraries (DELs) on live cells. This method connects extracellular ligand binding with intracellular biochemical transformation, thereby biasing the selection toward agonist identification. We have demonstrated the methodology with three membrane proteins: epidermal growth factor receptor (EGFR), thrombopoietin receptor (TPOR), and insulin receptor (INSR). A ∼30 million and a 1.033 billion-compound DEL were selected against these targets, and novel agonists with subnanomolar affinity and low micromolar cellular activities have been discovered. The INSR agonists activated the receptor by possibly binding to an allosteric site, exhibited clear synergistic effects with insulin, and activated the downstream signaling pathways. Notably, the agonists did not activate the insulin-like growth factor 1 receptor (IGF-1R), a highly homologous receptor whose activation may lead to tumor progression. Collectively, this work has developed an approach toward "functional" DEL selections on the cell surface and may provide a widely applicable method for agonist discovery for membrane proteins.

Encoding and display technologies for combinatorial libraries in drug discovery: The coming of age from biology to therapy

Drug discovery is a sophisticated process that incorporates scientific innovations and cutting-edge technologies. Compared to traditional bioactivity-based screening methods, encoding and display technologies for combinatorial libraries have recently advanced from proof-of-principle experiments to promising tools for pharmaceutical hit discovery due to their high screening efficiency, throughput, and resource minimization. This review systematically summarizes the development history, typology, and prospective applications of encoding and displayed technologies, including phage display, ribosomal display, mRNA display, yeast cell display, one-bead one-compound, DNA-encoded, peptide nucleic acid-encoded, and new peptide-encoded technologies, and examples of preclinical and clinical translation. We discuss the progress of novel targeted therapeutic agents, covering a spectrum from small-molecule inhibitors and nonpeptidic macrocycles to linear, monocyclic, and bicyclic peptides, in addition to antibodies. We also address the pending challenges and future prospects of drug discovery, including the size of screening libraries, advantages and disadvantages of the technology, clinical translational potential, and market space. This review is intended to establish a comprehensive high-throughput drug discovery strategy for scientific researchers and clinical drug developers.

Machine learning in preclinical drug discovery

Drug-discovery and drug-development endeavors are laborious, costly and time consuming. These programs can take upward of 12 years and cost US $2.5 billion, with a failure rate of more than 90%. Machine learning (ML) presents an opportunity to improve the drug-discovery process. Indeed, with the growing abundance of public and private large-scale biological and chemical datasets, ML techniques are becoming well positioned as useful tools that can augment the traditional drug-development process. In this Perspective, we discuss the integration of algorithmic methods throughout the preclinical phases of drug discovery. Specifically, we highlight an array of ML-based efforts, across diverse disease areas, to accelerate initial hit discovery, mechanism-of-action (MOA) elucidation and chemical property optimization. With advances in the application of ML across diverse therapeutic areas, we posit that fully ML-integrated drug-discovery pipelines will define the future of drug-development programs.

Evolution of chemistry and selection technology for DNA-encoded library

DNA-encoded chemical library (DEL) links the power of amplifiable genetics and the non-self-replicating chemical phenotypes, generating a diverse chemical world. In analogy with the biological world, the DEL world can evolve by using a chemical central dogma, wherein DNA replicates using the PCR reactions to amplify the genetic codes, DNA sequencing transcripts the genetic information, and DNA-compatible synthesis translates into chemical phenotypes. Importantly, DNA-compatible synthesis is the key to expanding the DEL chemical space. Besides, the evolution-driven selection system pushes the chemicals to evolve under the selective pressure, i.e., desired selection strategies. In this perspective, we summarized recent advances in expanding DEL synthetic toolbox and panning strategies, which will shed light on the drug discovery harnessing in vitro evolution of chemicals via DEL.