This is the first of a three-part series highlighting essential computational chemistry topics covered in the newly released Schrödinger Online Course bundle. Read on to learn how to get started with virtual screening and best practices to design an impactful ligand library.
In the vastness of chemical space, the number of possible drug-like molecules is estimated to be around 1060! Modern physics and machine-learning based applications enable the effective exploration of this near-infinite novel chemical space on a much larger scale than would ever be feasible without computation. Essential to this approach is a process called virtual screening, in which drug hunters search through a library of small molecules and screen for the ligand compounds, called “hits,” that are most likely to bind with a specific protein target.
Crafting a high-quality ligand library is essential for exploring chemical space and discovering new hits. While broad, general-purpose libraries exist, focusing on libraries customized for specific screening campaign goals and targets is more effective. This tailored approach can maximize results and accelerate the timeline for a given project.
The Fundamentals of Library Design
Designing a ligand library requires thoughtful selection and filtering processes to improve both relevance and quality. Here’s what makes a high-quality ligand library:
Increases Virtual Screening Success: A quality library is one that boosts the likelihood of identifying initial hits, even with moderate potency. These hits serve as starting points for further refinement, where potency and other characteristics are optimized in later stages.
Ensures Synthetic Feasibility: The hits should ideally be available for purchase or synthetically feasible. Understanding the limits of synthesizability in virtual compounds is vital to avoid costly setbacks. A compound might look promising in a virtual screen, but it’s crucial to confirm it can realistically be made in the lab.
Prioritizes Drug-Likeness and Predicted Properties: Filtering out non-druglike molecules early reduces the burden on computational resources and enhances the quality of screened compounds.
Excludes Undesirable Compounds: Libraries should avoid compounds with clear liabilities, such as REOS (rapid elimination of swill) or PAINS (pan-assay interference compounds), which can cause complications. Removing these structures before screening minimizes wasted resources on unsuitable candidates.
Emphasizing Diversity and Novelty in Chemical Space
Leveraging scaffold diversity and exploring a wide range of chemotypes increases the chances of finding unique, novel hits. Including varied chemotypes while excluding or setting aside patented compounds ensures a library tailored to the project’s requirements. This approach not only fosters innovation but also minimizes the risk of legal or intellectual property issues.
At times, it’s beneficial to design a library around known chemical matter for a target of interest, especially if there is prior knowledge about the target. Enriching your dataset with compounds that align with anticipated interactions can provide a strategic advantage, allowing you to focus on specific areas of chemical space that best suit your project’s needs.
Best Practices for Building Effective Ligand Libraries
In designing a ligand library for both virtual and lab-based screening techniques, it’s wise to filter by specific physical and chemical properties while maintaining sufficient diversity. Collaborate closely with medicinal chemists to understand campaign objectives, ensuring the library aligns with the target profile. Each ligand should represent a molecule you’re prepared to pursue if it turns up as a hit.
Applying aggressive filters and curating the library reduces costs and frustrations associated with unsuitable molecules. Additionally, running pilot screens with a subset of ligands allows for refinement before launching a full-scale virtual screening. While it’s impossible to guarantee library effectiveness before screening, these best practices increase the likelihood of successful outcomes and allow for adjustments later if needed.
Managing Virtual Screening Costs
Virtual screening involves substantial investments, including computational and human resources, analysis, and software licensing fees. Although virtual screening is faster than experimental wet lab-based methods, the vastness of chemical space makes it expensive and time-consuming. Proactive library design and pilot screens can cut down on these costs by avoiding post-screening processing of undesirable molecules.
Key Takeaways for Effective Ligand Libraries
The effectiveness of virtual screening workflows depends heavily on the libraries being screened. Even the most accurate model will struggle if the chosen ligands don’t align with the project’s objectives. For instance, it’s not recommended to use peptide-like compounds for a CNS project where smaller molecular weights are preferred.
Starting with commercially available libraries from reputable vendors often provides a strong foundation, though additional filtering and pilot screening remain crucial steps to ensure relevance.
Maintaining a high-quality ligand library can transform screening campaigns by minimizing time and resource investments. Thoughtful design not only enhances the likelihood of identifying promising hits but also reduces the risks associated with unsuitable molecules.
Next Steps
If you’re interested in diving deeper, Schrödinger has recently launched a new online certification course, Designing Quality Ligand Libraries, which combines in-depth learning with hands-on exercises to bring these concepts to life.
