The High-Throughput Screening Transformation in Modern Drug Development

Before HTS became integral to drug discovery, researchers relied on manual, hypothesis-driven methods. Each compound was tested individually against a biological target, often based on prior research or known biological activity. While this approach contributed to important pharmaceutical breakthroughs, it was inherently slow and lacked the scalability needed to meet the growing demands of modern drug development.^1,2

The need for greater efficiency spurred early automation efforts aimed at streamlining repetitive laboratory tasks such as pipetting, sample preparation, and data collection. These initial innovations laid the groundwork for what would become HTS by allowing researchers to test multiple compounds simultaneously. Early screening platforms focused on enzyme inhibition and receptor-binding assays — foundational tools that provided valuable insights into molecular interactions and biological activity.^1,3

As research needs grew, so did the sophistication of screening methods. The introduction of simple robotic systems significantly increased throughput, allowing scientists to process larger libraries of compounds in a fraction of the time previously required. This shift marked the beginning of a new era in drug discovery, one in which scale and automation would play increasingly central roles.²

The pharmaceutical industry’s adoption of HTS accelerated in the 1990s, driven by mounting pressure to reduce the time and cost associated with bringing new drugs to market. With the increasing complexity of drug development pipelines and the rising financial burden of research and development, companies recognized the need for more efficient, large-scale screening technologies.^4,5

This period was marked by significant technological advancements, including the introduction of automated liquid handling systems, sophisticated microplate formats, and high-speed detection technologies. The adoption of microplates — initially 96-well formats, later expanded to 384- and 1536-well configurations — enabled researchers to conduct thousands of assays simultaneously. Automation not only increased throughput but also improved precision and reproducibility while significantly reducing reagent consumption and overall costs.^4,6 Combinatorial chemistry expanded the scope of HTS by enabling the synthesis of diverse chemical libraries for screening.

Contemporary HTS platforms have advanced far beyond their early iterations, evolving into highly sophisticated systems that integrate miniaturization, multiplexing, and data-driven methodologies. Modern screening technologies allow researchers to conduct millions of tests simultaneously using only minimal amounts of reagents. High-density microplates, coupled with advanced detection systems, have dramatically increased the speed, accuracy, and scalability of screening operations.^7,8

Modern HTS benefits from the growing influence of big data analytics and computational biology. These tools allow researchers to refine screening strategies, optimize hit rates, and improve the overall efficiency of drug discovery pipelines. Moreover, automated high-throughput experimentation (HTE) has expanded the accessible chemical space by enabling the rapid synthesis and screening of diverse molecular scaffolds, increasing the likelihood of discovering novel therapeutics.^2,7

As HTS continues to evolve, its role in drug discovery has expanded beyond traditional small-molecule screening. Platforms are now increasingly used for phenotypic assays, genetic screenings, and biomarker discovery, further solidifying HTS as a central pillar of modern pharmaceutical research.^6,8

The Challenges HTS Solves in Drug Discovery

Before HTS became a cornerstone of drug discovery, researchers faced significant bottlenecks associated with manual compound testing. Traditional screening methods were labor-intensive, requiring researchers to test one compound at a time through repetitive, time-consuming assays. This not only limited the number of compounds that could be tested but also significantly slowed the pace of drug development, often extending the timeline for new therapies by years.²

The difficulty in screening large compound libraries efficiently was another major hurdle. With the growing complexity of disease biology and the need to explore vast chemical spaces for novel therapeutics, the limitations of manual methods became increasingly apparent. The sheer volume of potential compounds meant that only a small fraction could be realistically tested using conventional approaches, leaving many promising drug candidates unexplored.¹

HTS addressed these challenges by automating and miniaturizing the screening process. By enabling the simultaneous testing of thousands — or even millions — of compounds, HTS drastically reduced the time required to identify initial hits. Automation eliminated much of the manual labor involved in the process, while advances in assay technology allowed researchers to work with larger libraries without proportionally increasing costs or effort.³ This scalability has made HTS a transformative tool in modern drug discovery, allowing pharmaceutical companies to accelerate research timelines and improve overall productivity.

One of the most significant contributions of HTS is its ability to enhance hit identification and streamline lead optimization. Traditional drug discovery methods often struggled with the initial identification of bioactive compounds, particularly when screening against novel targets. The process of moving from an initial hit to a validated lead compound was slow and fraught with uncertainty, limiting the speed and success rate of early-stage drug development.⁹

HTS transformed this aspect of discovery by enabling the rapid identification of bioactive compounds from large chemical libraries. Automated platforms can process vast data sets in a fraction of the time it would take manual methods, allowing researchers to quickly identify molecules with promising biological activity. This acceleration of hit identification dramatically shortens the early stages of drug discovery, facilitating faster progress toward preclinical validation.⁹

Moreover, HTS supports streamlined hit-to-lead transitions by providing high-quality data that informs medicinal chemistry efforts. Researchers can quickly assess the structure–activity relationships (SAR) of hit compounds, allowing for rapid optimization of potency, selectivity, and pharmacokinetic properties. This ability to generate large volumes of relevant data accelerates the refinement of lead compounds, reducing the time and resources required to advance potential therapies into the development pipeline.³

One of the early criticisms of HTS was its propensity for generating false positives — compounds that appeared active during initial screening but failed to show efficacy upon further testing. Early iterations of HTS often struggled with assay interference, leading to unreliable results and wasted resources in downstream validation efforts.¹

Technological advancements have significantly addressed these issues by enhancing assay design, improving specificity, and increasing reproducibility. Modern HTS platforms now incorporate high-content screening (HCS), which captures more complex biological data, and label-free detection methods that reduce the risk of assay artifacts. These innovations have contributed to more reliable data, enabling researchers to focus on the most promising candidates.^1,6

Another strategy for reducing false positives involves the use of confirmatory screens and orthogonal assays. Confirmatory screens re-test hits using slightly modified conditions to verify their activity, while orthogonal assays assess compounds using entirely different detection methods to ensure biological relevance. This layered approach minimizes the risk of advancing false positives, conserving resources and improving the efficiency of the overall discovery process.²

HTS mitigates risk by identifying ineffective compounds early, reducing costs and improving the success rate of drug development pipelines.

Beyond eliminating ineffective compounds, HTS provides valuable data that supports more informed decision-making. The large datasets generated during HTS campaigns help guide medicinal chemistry efforts, inform SAR analyses, and refine target selection strategies. By identifying potential off-target effects early, HTS reduces the likelihood of late-stage failures, which are often financially and scientifically costly.⁴

In addition to improving the probability of success, HTS reduces downstream costs by ensuring that only the most promising compounds advance to further testing. This efficiency not only conserves resources but also accelerates development timelines, allowing companies to bring new therapies to market more quickly and cost-effectively.² As a result, HTS has become an essential tool for pharmaceutical companies seeking to streamline their discovery pipelines while mitigating financial and operational risks.

HTS relies on a range of assay formats to evaluate the biological activity of compounds quickly and accurately. The two primary categories — biochemical and cell-based assays — play distinct yet complementary roles in drug discovery. Biochemical assays typically focus on enzyme inhibition or receptor-binding interactions. Enzyme inhibition assays measure the ability of a compound to interfere with enzymatic activity, providing valuable insights for targeting specific metabolic pathways or enzymes associated with disease progression. Receptor-binding assays, on the other hand, assess how potential drug candidates interact with receptors, offering critical data for designing drugs that modulate signaling pathways.⁹

In recent years, phenotypic screening has gained traction as an alternative or complement to target-based assays. Unlike biochemical assays, phenotypic screening focuses on observing changes in cellular behavior, morphology, or function without prior knowledge of a specific molecular target. This unbiased approach has been particularly effective in identifying compounds with novel mechanisms of action, leading to breakthroughs in therapeutic areas such as oncology and neurodegenerative diseases.²

Another significant advancement in assay methodologies has been the development of label-free technologies, which eliminate the need for fluorescent or radioactive tags. Surface plasmon resonance (SPR) is one of the most widely adopted label-free methods, enabling real-time monitoring of molecular interactions with high sensitivity and specificity. This technique is particularly valuable for determining binding affinities and kinetics, which are critical for early-stage drug discovery.⁹

Fluorescence polarization assays also offer a powerful means of measuring molecular interactions, particularly in enzyme and receptor-binding studies. By detecting changes in the rotational motion of fluorescent-labeled molecules upon binding to a target, this technique provides a reliable and scalable solution for screening large compound libraries. The development of these advanced assay technologies has significantly improved the reliability and efficiency of HTS, allowing researchers to generate high-quality data at unprecedented scales.⁹

The introduction of automation and robotics has been one of the most transformative developments in HTS, enabling researchers to conduct experiments on an unprecedented scale while minimizing human error. Robotic liquid-handling systems have become standard tools in modern laboratories, automating processes, such as pipetting, reagent dispensing, and sample preparation. These systems not only increase throughput but also enhance precision and reproducibility by eliminating variability associated with manual techniques.⁷

Automation has also expanded the scope of accessible chemical space, allowing researchers to test more compounds than ever before. By integrating robotics with advanced data management systems, modern HTS platforms can handle vast compound libraries efficiently, significantly accelerating the pace of discovery. This ability to screen massive libraries quickly and accurately has made HTS a central pillar of pharmaceutical innovation, enabling the rapid identification of potential drug candidates from increasingly diverse chemical spaces.⁷

The integration of artificial intelligence (AI) and machine learning (ML) into HTS has ushered in a new era of data-driven drug discovery. AI algorithms are now being used to analyze large, complex data sets generated by HTS, uncovering patterns and correlations that might otherwise go unnoticed. This capability enhances the predictive power of screening campaigns, allowing researchers to identify promising hits more efficiently and with greater confidence.¹⁰

One of the most significant applications of AI in HTS is structure-based drug design. By using deep learning algorithms to model the interactions between drug candidates and their molecular targets, AI systems can predict binding affinities and prioritize compounds for further testing. This approach reduces the need for exhaustive physical screening, saving both time and resources while increasing the likelihood of identifying viable drug candidates.¹⁰

Virtual screening, powered by AI, further enhances the efficiency of HTS by narrowing down large compound libraries to a manageable number of high-potential candidates before physical testing begins. This pre-screening step allows researchers to focus their efforts on the most promising compounds, significantly accelerating the drug discovery process. By integrating AI with traditional HTS methodologies, pharmaceutical companies can optimize hit selection and streamline early-stage drug development.¹⁰

Automated HTE has dramatically expanded the chemical diversity accessible through modern drug discovery efforts. HTE platforms leverage robotics and advanced data analysis to explore vast chemical spaces systematically, enabling researchers to synthesize and test a wide array of molecular scaffolds rapidly.⁷

This expansion of chemical space is particularly important for identifying novel compounds with unique structures and mechanisms of action. By automating the synthesis and screening of diverse chemical libraries, HTE allows researchers to move beyond conventional chemical scaffolds and explore previously untapped molecular territories. This capability has led to the discovery of novel bioactive compounds that might have been overlooked using traditional screening methods.

In addition to increasing chemical diversity, automated experimentation enhances the efficiency of the discovery process by streamlining synthesis, purification, and testing workflows. Researchers can conduct multiple experiments simultaneously, generating large data sets that inform future screening campaigns and guide medicinal chemistry efforts. This systematic approach accelerates the identification of lead compounds, ultimately reducing the time and cost associated with bringing new therapies to market.

The integration of HTE with AI and big data analytics further amplifies its impact, allowing researchers to identify trends and correlations across vast data sets. This convergence of automation, data science, and chemistry has the potential to revolutionize drug discovery, making it possible to explore chemical spaces that were previously considered inaccessible or impractical to investigate.

While HTS has become a foundational technology in modern drug discovery, it is not the only approach available. Several alternative and complementary strategies have emerged to address limitations associated with HTS, such as the challenges of screening vast chemical libraries, the need for target validation, and the exploration of novel molecular mechanisms. These methodologies often work synergistically with HTS, enhancing its capabilities and providing additional layers of data that improve the overall efficiency and success rate of drug development.

Virtual screening has emerged as a powerful computational approach that complements traditional HTS workflows. This in silico method uses computational algorithms to evaluate large compound libraries based on their predicted ability to bind to a specific biological target. By simulating molecular interactions, virtual screening can rapidly narrow down potential candidates, prioritizing those with the highest likelihood of success for experimental validation.²

The integration of virtual screening with HTS creates significant efficiencies in drug discovery. Rather than physically testing every compound in a large library, researchers can first use virtual screening to identify a subset of high-potential candidates. These prioritized compounds can then undergo physical screening, streamlining the hit identification process and reducing the time, cost, and resources required for initial testing.²

In addition to accelerating the discovery process, virtual screening enhances predictive accuracy through structure-based drug design. By leveraging detailed structural data from X-ray crystallography, cryo-electron microscopy, or molecular modeling, researchers can refine virtual screening predictions to focus on compounds with the most promising binding affinities and pharmacokinetic properties. When combined with HTS, this approach creates a more targeted and efficient drug discovery strategy.

Fragment-based drug discovery (FBDD) offers an alternative strategy to traditional HTS by focusing on the identification of small molecular fragments that can bind to specific regions of a biological target. These fragments are typically much smaller than traditional drug-like molecules, allowing researchers to explore chemical space more efficiently and identify novel binding sites that larger compounds might miss.

Once initial fragments are identified, they can be chemically modified or linked together to develop compounds with higher potency and specificity. This iterative process allows for a more rational, structure-guided approach to drug design. Compared with HTS, FBDD typically involves screening much smaller libraries — often consisting of a few thousand fragments rather than millions of compounds — making the process more manageable and cost-effective for certain applications.

While HTS excels at identifying bioactive compounds from large chemical libraries, FBDD is particularly valuable for discovering novel scaffolds and targeting challenging biological structures, such as protein–protein interactions. Both methodologies offer distinct advantages, and in many cases, they are used in parallel to enhance the breadth and depth of drug discovery campaigns.

Functional genomics tools, particularly those based on CRISPR–Cas9 gene-editing technology, have introduced new dimensions to drug discovery. CRISPR-based functional screens allow researchers to systematically knock out, activate, or repress specific genes across the genome to identify their roles in disease pathways and therapeutic response.⁹

CRISPR screening complements HTS by providing a genetic validation platform for potential drug targets. For example, once HTS identifies a promising compound that affects a particular pathway, CRISPR screens can be used to determine which genes are essential for that pathway’s function. This helps confirm whether the observed biological activity is linked to the intended target, reducing the risk of off-target effects and increasing the likelihood of clinical success.⁹

Additionally, CRISPR-based screens are valuable for uncovering mechanisms of drug resistance, identifying genetic biomarkers of drug sensitivity, and exploring synthetic lethality relationships in cancer therapy. When integrated with HTS, these insights enable a more comprehensive understanding of drug–target interactions and support the development of more precise, targeted therapies.

HTS integrates automation and miniaturization, enabling rapid testing of vast chemical libraries with enhanced precision.

Automation plays a central role in reducing labor costs and minimizing human error. Robotic systems handle routine tasks such as liquid dispensing, sample preparation, and assay analysis with greater precision and speed than manual processes. This level of automation not only improves consistency and reproducibility but also significantly reduces the need for human intervention, allowing researchers to focus on more complex analytical tasks.⁴

Miniaturization further contributes to cost savings by decreasing the volume of reagents and samples required for each assay. The transition from traditional 96-well plates to higher-density formats has allowed researchers to conduct more assays simultaneously while using fewer resources. This reduction in reagent consumption not only cuts costs but also promotes more sustainable laboratory practices.⁴

Additionally, by consolidating various processes — such as compound dispensing, incubation, and detection — into a single automated workflow, HTS significantly reduces turnaround times. This efficiency enables faster decision-making and accelerates the overall pace of early-stage drug development, providing a clear economic advantage for research organizations.

Time efficiency is one of the most compelling benefits of HTS, particularly in the context of accelerating lead identification and preclinical testing. Traditional drug discovery methods can take years to move from hit identification to clinical trials, creating substantial delays in bringing new therapies to market. HTS drastically shortens this timeline by enabling the rapid screening of vast compound libraries in a matter of days or weeks.¹⁰

The integration of AI into HTS platforms has further enhanced time-saving capabilities. AI-driven predictive models can analyze complex data sets, identify patterns, and prioritize promising compounds for further testing. This capability reduces the need for exhaustive physical screening by narrowing the focus to compounds with the highest likelihood of success. By streamlining the hit-to-lead process, AI-supported HTS allows researchers to move from initial discovery to preclinical validation more quickly than traditional methods would permit.¹⁰

Accelerating lead identification directly impacts the overall time to market for new drugs. With the ability to quickly identify and optimize lead compounds, pharmaceutical companies can advance drug candidates into clinical trials more rapidly. This speed not only provides a competitive advantage but also ensures that potentially life-saving therapies reach patients sooner.

Beyond cost and time savings, HTS plays a critical role in mitigating the risks inherent in drug discovery and development. One of the primary ways HTS reduces risk is by enabling the early detection of off-target effects and potential toxicity. By screening large compound libraries across various biological targets, researchers can identify compounds that exhibit undesirable interactions early in the discovery process. Detecting these issues at an early stage prevents the advancement of compounds that are likely to fail in later, more costly phases of development.¹

HTS also improves the overall success rate of drug development pipelines by generating robust datasets that inform better decision-making. The wealth of data produced during screening campaigns allows researchers to assess the SAR of compounds, refine lead candidates, and optimize pharmacokinetic properties before entering preclinical testing. This data-driven approach helps identify potential liabilities early, reducing the likelihood of costly failures during clinical trials.¹

By integrating comprehensive screening with predictive analytics and confirmatory testing, HTS reduces both scientific and financial risk. This approach allows drug developers to allocate resources more effectively, minimize the chances of late-stage failures, and improve the overall probability of success in bringing new therapies to market.

As HTS continues to evolve, its future lies in integrating emerging technologies that enhance both the depth and precision of drug discovery. Innovations, such as 3D cell cultures, microfluidics, genetic screening, and AI, are set to expand the capabilities of HTS, pushing the boundaries of what is possible in early-stage drug development. These advancements not only promise to improve the physiological relevance of screening platforms but also accelerate the discovery of personalized therapies and novel drug targets.

Traditional HTS platforms have largely relied on two-dimensional (2D) cell cultures, which, while efficient and cost-effective, often fail to replicate the complex biological environments found in living organisms. This limitation can lead to discrepancies between in vitro findings and in vivo outcomes, reducing the predictive power of early-stage screens. The integration of three-dimensional (3D) cell cultures and organoids into HTS platforms aims to address this gap by providing more physiologically relevant models.

3D cultures better mimic the structural and functional complexity of human tissues, allowing researchers to observe how potential drug candidates behave in more realistic environments. Organoids, miniature organ-like structures derived from stem cells, offer an even more advanced model by replicating the architecture and functionality of specific human organs. These models are particularly valuable in personalized medicine, where patient-derived organoids can be used to assess individual responses to drug candidates, enabling tailored treatment strategies.

In the context of rare diseases, where patient populations are small and treatment options limited, 3D models offer a promising avenue for identifying therapies that might otherwise be overlooked. By incorporating these advanced models into HTS workflows, researchers can improve the predictive accuracy of preclinical testing, reducing the risk of failure in later development stages and paving the way for more targeted and effective treatments.

Microfluidic technologies, often referred to as lab-on-a-chip systems, represent another transformative advancement for HTS. These miniaturized platforms allow researchers to manipulate small volumes of fluids with high precision, enabling the execution of complex biological assays on a much smaller scale. This miniaturization reduces reagent consumption, lowers costs, and allows for high-throughput experimentation with increased speed and efficiency.

One of the key advantages of microfluidic systems is their ability to facilitate real-time analytics. By integrating sensors and detection systems directly onto the chip, researchers can monitor biological reactions as they occur, providing immediate feedback on compound activity, binding affinities, and toxicity profiles. This capability not only accelerates data collection but also improves the reproducibility and accuracy of screening results.

In addition, microfluidic platforms offer enhanced control over experimental conditions, enabling the simulation of physiological environments such as fluid flow, temperature gradients, and cellular interactions. These features make lab-on-a-chip technologies particularly well-suited for applications in personalized medicine, where patient-specific conditions can be modeled more accurately than in traditional HTS systems.

The emergence of CRISPR–Cas9 gene-editing technology has revolutionized functional genomics and opened new possibilities for genetic screening in drug discovery. CRISPR-based screens allow researchers to selectively activate, repress, or knock out specific genes, providing a powerful tool for understanding gene function and identifying potential drug targets.

When integrated with HTS, CRISPR-based genetic screening enhances the ability to validate targets and uncover mechanisms of drug resistance. For instance, large-scale CRISPR knockout screens can identify genes that contribute to disease progression or influence drug sensitivity, providing valuable insights that guide therapeutic development. These insights are particularly relevant in oncology, where gene-editing screens have been used to discover synthetic lethality relationships that can be exploited for cancer therapy.

CRISPR screening also plays a critical role in advancing personalized therapeutics and gene therapies. By identifying genetic vulnerabilities unique to individual patients, researchers can develop targeted treatments that address the specific molecular drivers of disease. As CRISPR technology continues to evolve, its integration into HTS platforms will further expand the potential for precision medicine and targeted drug development.

AI and ML are poised to redefine the capabilities of HTS by enhancing predictive modeling and streamlining the drug discovery process. Deep learning algorithms can analyze vast data sets generated by HTS campaigns, identifying patterns and correlations that are often invisible to traditional analytical methods. These predictive models can forecast how compounds will interact with specific biological targets, significantly reducing the need for exhaustive experimental testing.¹⁰

AI-driven virtual screening is particularly valuable for prioritizing compounds before physical testing begins. By simulating molecular interactions and predicting binding affinities, AI systems can focus HTS efforts on the most promising candidates, increasing the efficiency of the discovery pipeline. This approach accelerates hit identification, reduces costs, and improves the overall success rate of drug development initiatives.

Beyond efficiency gains, AI offers significant potential in personalized medicine. By incorporating patient-specific genetic and molecular data into predictive models, AI can help identify therapies tailored to an individual’s unique biological profile. This personalized approach not only enhances treatment efficacy but also reduces the risk of adverse reactions.

As AI technologies continue to advance, their integration with HTS platforms will enable faster, more accurate predictions, transforming early-stage drug discovery into a more targeted, data-driven process. This evolution promises to increase the speed and efficiency of drug development while expanding the potential for personalized and precision therapeutics.¹⁰

These future frontiers signal a transformative shift in HTS, as the integration of advanced biological models, microfluidic systems, genetic screening, and AI technologies pushes the boundaries of drug discovery. By embracing these innovations, researchers can accelerate the development of novel therapeutics, improve predictive accuracy, and bring personalized treatments closer to clinical reality.

As these technologies converge, HTS will remain at the forefront of pharmaceutical innovation, driving efficiency, reducing development costs, and mitigating the risks inherent in drug discovery. Its ongoing evolution will not only accelerate the development of new treatments but also enable a more personalized, data-driven approach to medicine—paving the way for breakthroughs that will redefine the future of healthcare.

Our parent company, That’s Nice, is committed to supporting the companies and innovators driving the next wave of pharma and biotech innovation. To celebrate That’s Nice’s 30th anniversary, Pharma’s Almanac is diving into 30 groundbreaking advancements, trends, and breakthroughs that have shaped the life sciences, highlighting the industry-defining milestones our agency has had the pleasure of growing alongside. Here’s to 30 years of innovation and the future ahead!

1. Pereirra, D. A., & Williams, J. A. “Origin and evolution of high throughput screening.” British Journal of Pharmacology. 157: 286–291 (2009).

2. Macarron, Ricardo et al. “The impact of high-throughput screening in biomedical research.” Nature Reviews Drug Discovery. 10(3), 188–195 (2011).

3. Mayr, L. M., & Bojanic, D. “Novel trends in high-throughput screening.” Drug Discovery Today. 14: 840–847 (2009).

4. Hajare, R et al. “Review on High-throughput screening is an approach to drug discovery.” International Journal of Pharmaceutical Sciences Review and Research. 21: 225–232 (2013).

5. Armstrong, Jason W. “A review of high-throughput screening approaches for drug discovery.” Target Discovery Institute, University of Oxford. Apr. 1999.

6. Szymanski, Pawel, Magdalena Markowicz, and Elzbieta Mikiciuk-Olasik. “Adaptation of high-throughput screening in drug discovery.” Assay and Drug Development Technologies. 10: 423–434 (2012).

7. Liver, E. “Expanding accessible chemical space through automated high-throughput experimentation.” Drug Target Review. 6 Apr. 2021

8. Aldewachi, Hasan, et al. “High-throughput screening platforms in the discovery of novel therapeutics for neurodegenerative disorders.” Bioengineering. 8: 30 (2021).

9. Kansagra, K, K Patel, V Shah, and A Desai. “A brief review of high throughput screening in drug discovery process.” Journal of Pharmaceutical Research International. 34: 47–55 (2022).

10. The Atomwise AIMS Program. “AI-driven drug discovery: Enhancing hit prediction through structure-based virtual screening.” Scientific Reports. 14: 7526 (2024).