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

Yanrui Suo,

Qian Xu, Zhaoping Xiong

et al.

Journal of Medicinal Chemistry, Journal Year: 2024, Volume and Issue: 67(21), P. 18969 - 18980

Published: Oct. 23, 2024

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

Language: Английский

Machine learning in preclinical drug discovery DOI

Denise B. Catacutan,

Jeremie Alexander,

Autumn Arnold

et al.

Nature Chemical Biology, Journal Year: 2024, Volume and Issue: 20(8), P. 960 - 973

Published: July 19, 2024

Language: Английский

Citations

44
Evolution of chemistry and selection technology for DNA-encoded library DOI Creative Commons
Peixiang Ma, Shuning Zhang, Qianping Huang

et al.

Acta Pharmaceutica Sinica B, Journal Year: 2023, Volume and Issue: 14(2), P. 492 - 516

Published: Oct. 11, 2023

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

Language: Английский

Citations

25
Encoding and display technologies for combinatorial libraries in drug discovery: The coming of age from biology to therapy DOI Creative Commons

Yu Fan,

Ruibing Feng, Xinya Zhang

et al.

Acta Pharmaceutica Sinica B, Journal Year: 2024, Volume and Issue: 14(8), P. 3362 - 3384

Published: April 10, 2024

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

Language: Английский

Citations

12
Machine-Learning-Based Data Analysis Method for Cell-Based Selection of DNA-Encoded Libraries DOI Creative Commons
Rui Hou,

Chao Xie,

Yuhan Gui

et al.

ACS Omega, Journal Year: 2023, Volume and Issue: 8(21), P. 19057 - 19071

Published: May 15, 2023

DNA-encoded library (DEL) is a powerful ligand discovery technology that has been widely adopted in the pharmaceutical industry. DEL selections are typically performed with purified protein target immobilized on matrix or solution phase. Recently, DELs have also used to interrogate targets complex biological environment, such as membrane proteins live cells. However, due landscape of cell surface, selection inevitably involves significant nonspecific interactions, and data much noisier than ones proteins, making reliable hit identification highly challenging. Researchers developed several approaches denoise datasets, but it remains unclear whether they suitable for cell-based selections. Here, we report proof-of-principle new machine-learning (ML)-based approach process datasets by using Maximum A Posteriori (MAP) estimation loss function, probabilistic framework can account quantify uncertainties noisy data. We applied dataset, where 7,721,415 compounds was selected against carbonic anhydrase 2 (CA-2) line expressing 12 (CA-12). The extended-connectivity fingerprint (ECFP)-based regression model MAP function able identify true binders structure-activity relationship (SAR) from datasets. In addition, regularized enrichment metric (known enrichment) could be calculated directly without involving specific model, effectively suppressing low-confidence outliers enhancing signal-to-noise ratio. Future applications this method will focus de novo

Language: Английский

Citations

18
Agonist Discovery for Membrane Proteins on Live Cells by Using DNA-encoded Libraries DOI
Yiran Huang, Rui Hou, Fong Sang Lam

et al.

Journal of the American Chemical Society, Journal Year: 2024, Volume and Issue: 146(35), P. 24638 - 24653

Published: Aug. 22, 2024

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

Language: Английский

Citations

5
A Target Class Ligandability Evaluation of WD40 Repeat-Containing Proteins DOI Creative Commons
Suzanne Ackloo, Fengling Li,

Magda Szewczyk

et al.

Journal of Medicinal Chemistry, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 4, 2024

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 systematic approach to scale up the and characterization small molecule ligands for WD40 repeat (WDR) family. We developed comprehensive suite protocols production, crystallography, biophysical, biochemical, cellular assays. A pilot hit-finding campaign using DNA-encoded chemical library selection followed by machine learning (DEL-ML) predict from virtual libraries yielded first-in-class, drug-like 7 16 WDR domains screened, thus demonstrating broader ligandability WDRs. This study establishes template evaluation family wide provides an extensive resource biochemical tools, knowledge, discover potential therapeutics this highly disease-relevant, but underexplored target class.

Language: Английский

Citations

4
DNA-encoded Library Machine Learning Applications DOI

Eric A. Sigel

Royal Society of Chemistry eBooks, Journal Year: 2025, Volume and Issue: unknown, P. 17 - 40

Published: Feb. 21, 2025

Machine learning (ML) has begun to realize its promise in many domains the last several years. While small molecule drug discovery lagged comparison other areas, developments computing capabilities, data generation, and algorithms have enabled significant progress prediction. DNA-encoded libraries (DELs) represent an efficient way generate quantity of required for effective model building, providing a mechanism protein-target specific prediction with economics that permit individual organizations operate. DEL-based machine (DEL-ML) been demonstrated work variety targets continues expand usage industry approaches reported. With this initial success, number challenges considerations faced by DEL-ML practitioner identified including denoising DEL data, choice ML algorithm, hyperparameters representations, need relevant metrics assessment, particularly given high resource time costs testing predictions. In order fully potential DEL-ML, key improvements infrastructure broad availability are needed.

Language: Английский

Citations

0
Solid-phase DNA-encoded Library Technology DOI Creative Commons
Juan Hu,

John P. Burdick,

Brian M. Paegel

et al.

Royal Society of Chemistry eBooks, Journal Year: 2025, Volume and Issue: unknown, P. 41 - 62

Published: Feb. 21, 2025

Solid-phase DNA-encoded library (DEL) technology introduces advanced activity-based screening capabilities by virtue of its “one-bead-one-compound” (OBOC) format. In this review, we first describe the design and construction so-called “OBOC-DELs.” We then explore engineering a microfluidic platform that integrates automates high-throughput bead-based screening, highlighting examples fluorescence-based functional assay development miniaturization to droplets. Additionally, detail statistical framework OBOC-DEL experimental data interpretation. Finally, summarize numerous applications have spawned since technology’s inception, including biochemical activity, dose-response, cellular competition binding affinity, pharmacokinetic properties. Looking forward, there are likely further opportunities employ synthesis strategies other encoded modalities.

Language: Английский

Citations

0
Assays of DNA-encoded Libraries Against Protein Targets on and Within Living Cells DOI

Siavash Shahbazi Nia,

Casey J. Krusemark

Royal Society of Chemistry eBooks, Journal Year: 2025, Volume and Issue: unknown, P. 63 - 86

Published: Feb. 21, 2025

Assay platforms available for DNA-encoded chemical libraries (DELs) are largely limited to an in vitro selection assay binding a biochemical pure protein on solid support. Extending DEL assays proteins the cell surface and within live cells offers ability targets that cannot be reconstituted biochemically more physiologically relevant state. Significant challenges exist hinder cellular application of DELs. In this review, we summarise various approaches have been applied date enable against both cells. We discuss benefits limitations these how they address unique assays. explore potential molecular discovery from varying complexity. highlight some molecules discovered successfully with lastly offer outlook future.

Language: Английский

Citations

0
Drug discovery technologies–Current and future trends DOI

Mark C. Noe,

Claire M. Steppan,

Andrea D. Weston

et al.

Elsevier eBooks, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 1, 2025

Language: Английский

Citations

0