High-throughput Screening Method Research Articles

Abstract B001: CircRNA-derived neoantigen identification in cancer vaccine development

Abstract Immunotherapy has been a popular topic in combating cancer for its great performance in specificity, versatility and durability recently. Cancer vaccines targeting neoantigens have shown clinical efficacy against cancer, eliciting long-lasting immune response and offering a more selective therapeutic approach with less side effects. However, rare mutation rates and low expression levels of neoantigen make it challenging to identify suitable neoantigens. Traditionally, researchers focus on the canonical mRNA transcriptome and proteome to find potential targets. Recent studies suggest that exploring the non-canonical transcriptome and proteome, particularly circular RNAs (circRNAs), can expand the search. CircRNAs, endogenously formed by the unique back-splicing events during pre-RNA processing, can serve as templates for protein synthesis. Some of these non-canonical peptides are specifically expressed in cancer cells, making circRNAs a valid source for neoantigen discovery. We have developed high-throughput circRNA reporter screening methods and mass spectrometry-based peptidomic workflow to pinpoint translating circRNAs and their derived peptides. Using these techniques, we have identified circRNA-derived neoantigens in breast cancer (BC) cell lines that can be used as possible targets for cancer vaccines. These circRNA-derived neoantigens are immunogenic and can trigger dendritic cell (DC) maturation and T cell activation in vitro. Next, we will test these findings by transferring to in vivo mouse system. We will assess cancer vaccine efficacy by targeting the neoantigens using syngeneic mouse tumor models. Our long- term goal is to apply the circRNA neoantigen identification technology to BC patient samples to facilitate cancer vaccine development and improve cancer immunotherapy. Our work will demonstrate the potential of circRNA-derived non canonical neoantigens in cancer vaccines, serving as a crucial first step to prove their effectiveness in therapy. By expanding our search beyond the traditional transcriptome and proteome, we will explore new targets and transform the neoantigen-based treatment. This innovative effort shifts the current understanding of neoantigen sources and opens up new possibilities for targeted cancer therapy strategies. Citation Format: Yi-Cian Zheng, Chun-Kan Chen. CircRNA-derived neoantigen identification in cancer vaccine development [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: RNAs as Drivers, Targets, and Therapeutics in Cancer; 2024 Nov 14-17; Bellevue, Washington. Philadelphia (PA): AACR; Mol Cancer Ther 2024;23(11_Suppl):Abstract nr B001.

A yeast surface display platform for screening of non-enzymatic protein secretion in Kluyveromyces lactis.

Enhancing the secretion of recombinant proteins, particularly non-enzymatic proteins that predominate in food and pharmaceutic protein products, remains a significant challenge due to limitations in high-throughput screening methods. This study addresses this bottleneck by establishing a yeast surface display system in the food-grade microorganism Kluyveromyces lactis, enabling efficient display of model target proteins on the yeast cell surface. To assess its potential as a universal high-throughput screening tool for enhanced non-enzymatic protein secretion, we evaluated the consistency between protein display levels and secretion efficiency under the influence of various genetic factors. Our results revealed a strong correlation between these two properties. Furthermore, screening in a random mutagenesis library successfully identified a mutant with improved secretion. These findings demonstrate the potential of the K. lactis surface display system as a powerful and universal tool for high-throughput screening of strains with superior non-enzymatic protein secretion capacity. We believe this study could pave the way for efficient large-scale production of heterologous food and therapeutic proteins in industries. KEY POINTS: • A YSD (yeast surface display) system was established in Kluyveromyces lactis • This system enables high-throughput screening of non-enzymatic protein secretion • This technology assists industrial production of food and therapeutic proteins.

Drug Repurposing against Novel Therapeutic Targets in Plasmodium falciparum for Malaria: The Computational Perspective.

Malaria remains one of the most challenging tropical diseases. Since malaria cases are reportedly alarming in terms of infections and mortality, urgent attention is needed for addressing the issues of drug resistance in falciparum malaria. High throughput screening methods have paved way for rapid identification of anti-malarial. Furthermore, drug repurposing helps in shortening the time required for drug safety approvals. Hence, discovery of new antimalarials by drug repurposing is a promising approach for combating the disease. This article summarizes the recent computational approaches used for identifying novel antimalarials by using drug target interaction tools followed by pharmacokinetic studies.

Multi-residue analytical method development and dietary exposure risk assessment of 345 pesticides in mango by LC-Q-TOF/MS

Mango is attracting increasing global attention for its superior flavor and nutritional benefits. However, the use of pesticides has resulted in a threat to its quality, safety, and human health due to their inevitable residue. There is an urgent need for advanced technologies that can efficiently and sensitively analyze a wide range of pesticides. In this study, a high-throughput screening and quantification method for 345 pesticides in mango was established by applying a modified QuEChERS method combined with liquid chromatography-quadrupole-time-of-flight mass spectrometry (LC-Q-TOF/MS). A response surface methodology (RSM) optimized the critical parameters for extracting various pesticides in 1% acetic acid-acetonitrile (v/v) using MgSO4, PSA, C18, and GCB as the adsorbents. The proposed method was successfully validated and applied to the quantitative detection of pesticide residues in 73 mango samples from Hainan. The study found pesticide residues in 95.9% of mango samples, of which 3 samples had residue concentrations in excess of the Chinese maximum residue limit. A total of 24 pesticides were detected, with detection percentages ranging from 1.4% (Thiabendazole) to 41.1% (Pyraclostrobin). Considering the possible health risks related to pesticide residues, risk assessment of human exposure to pesticides via intake of mango was evaluated. The acute and chronic dietary exposure risks related to pesticide residues in mango were accepted. In conclusion, this study provides a powerful platform for monitoring multiple pesticides in mango. First, it can promote the rational use of pesticides, and second, it can help consumers to assess the risk of dietary exposure to pesticide residues.

Restoring Colistin Sensitivity in Multidrug-Resistant Pathogenic E. coli Using Cinacalcet Hydrochloride

Restoring colistin’s efficacy is crucial in addressing the resistance crisis of colistin. This study utilized a high-throughput screening method to identify 43 compounds from 800 FDA-approved drugs that exhibited significant antibacterial effects when combined with colistin. Among these, cinacalcet hydrochloride (CH) was selected for its potential synergistic effect with colistin against multidrug-resistant (MDR) E. coli strains, including mcr-1-positive strains. A series of experiments revealed that the combination of CH and colistin showed strong synergy, especially in mcr-1-positive strains, restoring colistin sensitivity. The combination significantly inhibited bacterial growth and reduced CFU counts more effectively than either drug alone. Additionally, CH and colistin together significantly inhibited biofilm formation and eradicated existing biofilms, as visualized through confocal microscopy. Mechanistic studies showed that the combination increased bacterial membrane permeability and disrupted membrane integrity. The treatment also elevated extracellular ATP release and ROS production, indicating oxidative stress-induced bacterial death. Safety evaluations showed that the combination did not increase toxicity in host cells. Finally, animal models further validated the combination’s efficacy. Overall, this study showed that the combination of colistin and CH significantly restored colistin sensitivity in mcr-1-positive E. coli, revealing their synergistic antibacterial mechanism involving membrane damage and oxidative stress, with promising clinical applications.

The 'Omics Revolution in CHO Biology: Roadmap to Improved CHO Productivity.

Chinese hamster ovary (CHO) cell physiology understanding has advanced very rapidly in the past few years with incredible improvements in long-read sequencing, improved resolution, and increased computational power. Multiple parental lines have been sequenced and the resultant pan-genome can be leveraged to increase our understanding of the diverse pathways CHO cells can take to get high-productivity phenotypes. The same improvements in workflows have complemented transcriptomic studies. Microfluidics and label-free innovations have further increased the sensitivity and accuracy of proteomic methods, while also making proteomics more accessible. In this 'omics era, high-throughput screening methods, sophisticated informatic tools, and models continually drive major innovations in cell line development and process engineering. This review describes the various recent achievements in 'omics techniques and their application to improve recombinant protein expression from CHO cell lines.

Selection of Nucleotide-Encoded Mass Libraries of Macrocyclic Peptides for Inaccessible Drug Targets.

Technological advances and breakthrough developments in the pharmaceutical field are knocking at the door of the "undruggable" fortress with increasing insistence. Notably, the 21st century has seen the emergence of macrocyclic compounds, among which cyclic peptides are of particular interest. This new class of potential drug candidates occupies the vast chemical space between classic small-molecule drugs and larger protein-based therapeutics, such as antibodies. As research advances toward clinical targets that have long been considered inaccessible, macrocyclic peptides are well-suited to tackle these challenges in a post-rule of 5 pharmaceutical landscape. Facilitating their discovery is an arsenal of high-throughput screening methods that exploit massive randomized libraries of genetically encoded compounds. These techniques benefit from the incorporation of non-natural moieties, such as non- proteinogenic amino acids or stabilizing hydrocarbon staples. Exploiting these features for the strategic architectural design of macrocyclic peptides has the potential to tackle challenging targets such as protein-protein interactions, which have long resisted research efforts. This Review summarizes the basic principles and recent developments of the main high-throughput techniques for the discovery of macrocyclic peptides and focuses on their specific deployment for targeting undruggable space. A particular focus is placed on the development of new design guidelines and principles for the cyclization and structural stabilization of cyclic peptides and the resulting success stories achieved against well-known inaccessible drug targets.

High-throughput optimization of organic carbon provision strategies enables enhanced arachidonic acid production in novel microalgae

BackgroundMicroalgae are potential sustainable resources for the production of value-added chemicals that can be used as biofuels, pharmaceuticals, and nutritional supplements. Arachidonic acid (ARA), a omega-6 fatty acid, plays a crucial role in infant development and immune response, and can be used in cosmetics and pharmaceuticals. Demand for industrial-scale ARA production is continuously increasing because of its broad applicability. To address this demand, there has been a significant shift towards microorganism-based ARA production. To accelerate large-scale ARA production, it is crucial to select suitable strains and establish optimal culture conditions.ResultsHere, we isolated a novel microalga Lobosphaera incisa CFRC-1, a valuable strain that holds promise as a feedstock for ARA production. Optimal cultivation conditions were investigated using a high-throughput screening method to enhance ARA production in this novel strain. Out of 71 candidates, four organic carbon substrates were identified that could be utilized by L. incisa CFRC-1. Through flask-scale verification, fructose was confirmed as the optimal organic carbon substrate for promoting microalgal growth, total lipid accumulation, and ARA production. Subsequently, we investigated appropriate substrate concentration and cultivation temperature, confirming that the optimal conditions were 30 g L− 1 of fructose and 27 ℃ of temperature. Under these optimized conditions, biomass and ARA production reached 13.05 ± 0.40 g L− 1 and 97.98 ± 7.33 mg L− 1, respectively, representing 9.6-fold and 5.3-fold increases compared to the conditions before optimization conditions. These results achieved the highest biomass and ARA production in flask-scale cultivation, indicating that our approach effectively improved both production titer and productivity.ConclusionsThis study presents a novel microalgae and optimized conditions for enhancing biomass and ARA production, suggesting that this approach is a practical way to accelerate the production of valuable microalgae-based chemicals. These findings provide a basis for large-scale production of ARA-utilizing microalgae for industrial applications.

Maltose gradient-induced biosensor-based high-throughput screening for directed evolution of maltogenic amylase from Bacillus stearothermophilus

Maltogenic amylase is a starch-hydrolyzing enzyme commonly used in bread baking and high-concentration maltose syrup production. However, low catalytic activity limits its industrial application. Improving catalytic activity based on molecular modification and directed evolution requires a High-Throughput Screening (HTS) method. In this study, a maltose gradient-induced (MaGI) biosensor was designed and applied for the directed evolution of maltogenic amylase AmyM, showing a good positive correlation between enzyme activity and fluorescence. The MaGI biosensor detected maltose and maltogenic amylase activity efficiently and specifically. Two mutants, Q440N and S442N/Q661L, were identified through the screening of 3000 mutants using the MaGI biosensor, showing a significant increase in catalytic activity of 35.56 % and 24.51 %, respectively, compared to the wild-type. Meanwhile, the t1/2 of Q440N and S442N/Q661L at 60 °C increased by 58.53 % and 66.66 %, respectively. In industrial applications, the enhancement of catalytic activity and stability is conducive to improving production efficiency and reducing costs. MD simulation has found that when modifying multidomain enzymes, distal mutations can enhance catalytic activity. In conclusion, the developed MaGI biosensor is a promising tool for high-throughput and specific detection of maltose.

Using Reporter Gene Assays to Screen and Identify Chemical Compounds that Modulate Estrogen Receptor Activity.

Estrogen receptor alpha (ERα) is a nuclear receptor that is expressed mainly in the breast, uterus, and ovary, among several other organs. ERα plays important roles in reproduction, mammary gland formation, and glucose homeostasis. Disruption of ERα may result in adverse outcomes, such as cancer, impaired fertility, and abnormal fetal growth. Therefore, identifying compounds that modulate ERα is of great interest due to their potential-endocrine disrupting capability and pharmaceutical applications. To rapidly test tens of thousands of compounds, high-throughput screening assays are essential. Here, we describe high-throughput screening methods, including plating and treatment of cells in 384-well and 1536-well plates and analysis of the resulting data. The two cell lines used, MCF7-VM7Luc4E2 and HEK293-ERα-bla, have been described previously. MCF7-VM7Luc4E2 cells are a stable luciferase reporter gene cell line expressing full-length endogenous estrogen receptor in the MCF7 cell line background, and HEK293-ERα-bla cells stably express an ERα ligand-binding domain/GAL4 DNA-binding domain fusion regulating a UAS β-lactamase reporter gene. These cell lines can be used to identify and confirm ERα modulators. Published 2024. This article is a U.S. Government work and is in the public domain in the USA. Current Protocols published by Wiley Periodicals LLC. Basic Protocol 1: Establishment of a high-throughput ERα reporter gene assay with luminescence readout to identify activators and inhibitors of estrogen receptor α Basic Protocol 2: Use of an orthogonal assay with fluorescence readout to confirm potential estrogen receptor activators or inhibitors.

Simulation screening of umami peptides in the microbiota of soy sauce based on molecular docking and molecular dynamics

Simulation screening and molecular simulation techniques are gradually being applied to the study of bioactive peptides. In this study, we explored and simulated the screening process using metagenomic sequencing data of microorganisms from soy sauce. An LSTM-based screening model was constructed to identify the top 10 scoring peptides from a sample dataset of 19,335 raw peptides, primarily derived from Bacillus and Weissella confusa. Through homology modeling, molecular docking, and molecular dynamics simulations, Ser, Asp, Asn, Glu, and Gln were identified as the main amino acids involved in ligand-receptor binding. Hydrogen bonding, van der Waals forces, and hydrophobic interactions were revealed as the primary forces mediating receptor and ligand binding. The umami thresholds of ISWCFTY and QISPYRRI were verified through wet experiments to be 0.13 mg/mL and 0.11 mg/mL, respectively, exhibiting high umami intensity. Overall, this study presents a high-throughput screening method for predicting umami peptides, offering cost reduction, simplified steps, and saved manpower.

Development of a multi-analysis model using an epithelial-fibroblast co-culture system as an alternative to animal testing

The evaluation of respiratory chemical substances has been mostly performed in animal tests (OECD TG 403, TG 412, TG 413, etc.). However, there have been ongoing discussions about the limited use of these inhalation toxicity tests due to differences in the anatomical structure of the respiratory tract, difficulty in exposure, laborious processes, and ethical reasons. Alternative animal testing methods that mimic in vivo testing are required. Therefore, in this study, we established a co-culture system composed of differentiated epithelial cells under an air-liquid interface (ALI) system in the apical part and fibroblasts in the basal part. This system was designed to mimic the wound-healing mechanism in the respiratory system. In addition, we developed a multi-analysis system that simultaneously performs toxicological and functional evaluations. Several individual assays were used sequentially in a multi-analysis model for pulmonary toxicity. Briefly, cytokine analysis, histology, and cilia motility were measured in the apical part, and cell migration and gel contraction assay were performed by exposing MRC-5 cells to the basal culture. First, human airway epithelial cells from bronchial (hAECB) were cultured under air-liquid interface (ALI) system conditions and validated pseudostratified epithelium by detecting differentiation-related epithelial markers using Transepithelial Electrical Resistance (TEER) measurement, Hematoxylin and Eosin (H&E) staining, and immunocytochemistry (ICC) staining. Afterward, the co-culture cells exposed to Transforming growth factor-beta 1 (TGF-β1), a key mediator of pulmonary fibrosis, induced significant toxicological responses such as cytotoxicity, cell migration, and gel contraction, which are wound-healing markers. In addition, cilia motility in epithelial cells was significantly decreased compared to control. Therefore, the multi-analysis model with a 3D epithelial-fibroblast co-culture system is expected to be useful in predicting pulmonary toxicity as a simple and efficient high-throughput screening method and as an alternative to animal testing.

High-Throughput Screening of Antioxidant Drug Candidates from Natural Antioxidants with a "Zero" Intrinsic Fluorescence Peroxynitrite Sensing Precursor.

The fluorescence high-throughput screening method is of importance for new antioxidant drug candidate discovery for the treatment of serious hepatorenal syndrome, which displayed an obvious upregulated peroxynitrite level. However, most of the current ONOO- probes possessed incomplete fluorescence quenching efficiency, which can result in non-negligible probe inherent fluorescence. Hence, we utilized the probe conjugated structure disruption strategy to construct hydrogenation phosphorus-substituted rhodamine (H-PRh) with "zero" probe inherent fluorescence character. Based on the precursor, a series of natural products were screened for identifying antioxidant drug candidates. Luteolin was screened out by activating the Sirt1-Nrf2-HO-1 signaling pathway to regulate the accumulation of ONOO- in the hepatorenal syndrome. Overall, the "zero" probe inherent fluorescence ONOO- sensor constructed here applies for a promising and versatile toolbox for illuminating the ONOO--related pathological process in the hepatorenal syndrome. Besides, this strategy of constructing highly sensitive sensors could serve as a valuable reference for further fluorescent probes.

A High-Throughput Screening Approach for Evaluating the Binding Affinity of Environmental Pollutants to Nuclear Receptors.

Increasing levels of compounds have been detected in the environment, causing widespread pollution and posing risks to human health. However, despite their high environmental occurrence, there is very limited information regarding their toxicological effects. It is urgent to develop high-throughput screening (HTS) methods to guide toxicological studies. In this study, a receptor-ligand binding assay using an HTS system was developed to determine the binding potency of environmental pollutants on nuclear receptors. The test is conducted using a microplate reader (i.e., a 96-well plate containing various chemicals) by measuring the fluorescence polarization (FP) of a specific fluorescent probe. This assay consists of four parts: the construction and transformation of recombinant vectors, the expression and purification of the receptor protein (ligand-binding domain), receptor-probe binding, and competitive binding of chemicals with the receptor. The binding potency of two environmental pollutants, perfluorooctanesulfonic acid (PFOS) and triphenyl phosphate (TPHP), with peroxisome proliferator-activated receptor gamma (PPARγ) was determined to illustrate the assay procedure. Finally, the advantages and disadvantages of this method and its potential applications were also discussed.

A High-Throughput Visual Screen for the Directed Evolution of Cβ-stereoselectivity of L-threonine Aldolase.

L-Threonine aldolase (L-TA) is a pyridoxal phosphate-dependent enzyme that catalyzes the reversible condensation of glycine and aldehydes to form β-hydroxy-α-amino acids. The combination of directed evolution and efficient high-throughput screening methods is an effective strategy for enhancing the enzyme's catalytic performance. However, few feasible high-throughput methods exist for engineering the Cβ-stereoselectivity of L-TAs. Here, we present a novel method of screening for variants with improved Cβ-stereoselectivity; this method couples an L-threo-phenylserine dehydrogenase, which catalyzes the specific oxidation of L-threo-4-methylsulfonylphenylserine (L-threo-MTPS), with the concurrent synthesis of NADPH, which is easily detectable via 340-nm UV absorption. This enables the visual detection of L-threo-MTPS produced by L-TA through the measurement of generated NADPH. Using this method, we discover an L-TA variant with significantly higher diastereoselectivity, increasing from 0.98 % de (for the wild-type) to 71.9 % de.

Clear zone formation in microdroplets for high-throughput screening for lactic acid bacteria.

Droplet microfluidic-based technology is a powerful tool for biotechnology, and it is also expected that it will be applied to culturing and screening methods. Using this technology, a new high-throughput screening method for lactic acid bacteria was developed. In this study, the conventional culture of lactic acid bacteria that form clear zones on an agar medium was reproduced in water-in-oil droplets, and only the droplets in which lactic acid bacteria grew were collected one by one. Using this method, the specific recovery of Lactiplantibacillus plantarum from a mixture of L. plantarum and Escherichia coli and the acquirement of lactic acid bacteria from an environmental sample were successful. This method could be applied to various conventional screening methods using the clear zone as a microbial growth indicator. This has expanded the possibilities of applying droplet microfluidic-based technology to microbial cultivations.

Acrolein scavengers and detoxification: From high-throughput screening of flavonoids to mechanistic study of epigallocatechin gallate

Acrolein (ACR) is a widespread, highly toxic substance that poses significant health risks. Flavonoids have been recognized as effective ACR scavengers, offering a possible way to reduce these risks. However, the lack of specific high-throughput screening methods has limited the identification of ACR scavengers, and their actual detoxifying capacity on ACR remains unknown. To address this, we developed a high-throughput screening platform to assess the ACR scavenging capacity of 322 flavonoids. Our results showed that 80.7 % of the flavonoids could scavenge ACR, but only 34.4 % exhibited detoxifying effects in an ACR-injured QSG7701 cell model. Some flavonoids even increased toxicity. Structure-activity relationship (SAR) analysis indicated that galloyl and pyrogallol units enhance scavenging but worsen ACR-induced cytotoxicity. Further investigation revealed that epigallocatechin gallate (EGCG) could exacerbate ACR-induced redox disorder, leading to cell apoptosis. Our findings provide crucial data on the scavenging and detoxifying capacities of 322 flavonoids, highlighting that ACR scavengers might not mitigate ACR-induced toxicity and could pose additional safety risks.

High-throughput screening of green amino acid and surfactant mixtures with high corrosion inhibition efficiency: Experimental and modelling perspectives

A high-throughput screening method is developed for rapid evaluation of corrosion inhibitors. By employing a fluorometric technique, this method quickly assesses the concentration of iron ions converted from corrosion products in the reaction pool to rapidly evaluate the corrosion inhibition performance of amino acid and surfactants. Utilizing this high-throughput screening platform, the corrosion inhibition effects of L-cysteine mixed with quaternary ammonium salts of different alkyl chain length were evaluated, and their optimal mixing ratios were determined. Through corrosion experiments and molecular simulations, the synergistic mechanisms of the inhibitor mixtures are revealed.

Indigo production identifies hotspots in cytochrome P450 BM3 for diversifying aromatic hydroxylation.

Evolution of P450 BM3 is a topic of extensive research, but screening the various substrate/reaction combinations remains a time-consuming process. Indigo production has the potential to serve as a simple high-throughput method for reaction screening, as bacterial colonies expressing indigo (+) variants can be visually identified via their blue phenotype. Indigo (+) single variants, indigo (-) single variants and a combinatorial library, containing mutations that enable the blue phenotype, were screened for their ability to hydroxylate a panel of 12 aromatic compounds using the 4-aminoantipyrine colorimetric assay. Recombination of indigo (+) single variants to create a multiple-variant library is a particularly useful strategy, as all top performing P450 BM3 variants with high hydroxylation activity were either indigo (+) single variants or contained multiple substitutions. Furthermore, active variants, as determined using the 4-AAP assay, were further characterized and several variants were identified that gave more than 90% conversion with 1,3-dichlorobenzene and predominantly formed 2,6-dichlorophenol; other variants showed significant substrate selectivity. This supports the hypothesis that substitution at positions that enable the indigo (+) phenotype, or hotspot residues, is a general mechanism for increasing aromatic hydroxylation activity. Overall, this research demonstrates that indigo (+) single variants, identified via colorimetric colony-based screening, may be recombined to generate a multiply-substituted variant library containing many variants with high aromatic hydroxylation activity. The combination of colony-based screening and other screening assays greatly accelerates enzyme engineering, as readily-identified indigo (+) single variants can be recombined to create a library of active multiple variants without extensive screening of single variants.

Chemical-guided screening of top-performing metal–organic frameworks for hydrogen storage: An explainable deep attention convolutional model

Metal-organic framework (MOF)-based adsorptive hydrogen storage holds promise for enhancing the sustainable design of hydrogen storages by enhancing the usable volumetric (UV) and gravimetric (UG) capacities. However, the extensive number of MOFs poses a challenge in the search for optimal materials owing to the lack of an efficient and interpretable high-throughput screening method. This study introduces an explainable artificial intelligence (XAI) framework to expedite the discovery of high-capacity hydrogen adsorbents by predicting the UV and UG capacities using an attention densely connected convolutional (ADCC) network. A new hybrid dataset with various operating conditions and comprising 24 physical–chemical descriptors, such as void fraction (VF) and metal percentage (MP), was utilized to develop the ADCC model. The explainable ADCC model demonstrated superior predictive performance for the UV and UG capacities, with R2 values of 0.9886 and 0.9982, respectively. The inclusion of the chemical descriptors MOFs enhanced the prediction accuracy of the ADCC model. The XAI analysis showed that VF and MP dominated physical and chemical descriptors, respectively, for UV and UG. Consequently, the ADCC model identified EFAYIU—a real MOF—as a promising hydrogen storage material with UV and UG capacities of 51.55 g H2/L and 11.37 wt%, respectively, surpassing the current materials for hydrogen storage. Additionally, the identified EFAYIU was validated based on molecular simulations, confirming the high hydrogen adsorptive capacities obtained by the ADCC model. Thus, the proposed AI-based high-throughput screening method enables the rapid discovery of high-performance MOFs for sustainable hydrogen storage.