3D in Silico Screening Solutions to Advance Drug Discovery Project Development
Scientist
on
June 19, 2025
This blog post was written by Cresset, a registered Scientist.com supplier that delivers solutions to optimize chemistry R&D innovation and efficiency, enabling customers to accelerate the development of better treatments to improve the health of humans, animals and the environment. Their services are available on the Scientist.com marketplace.
Introduction
The phrase virtual screening covers a wide range of techniques, each with their own strengths, weaknesses and starting requirements. This means there is not a single virtual screening method that is universally suitable for every project, and furthermore, only by understanding a project and its goals can an appropriate determination be made for which is the right methodology to pursue.
Typically, virtual screening is split into two classifications: structure-based and ligand-based. The former uses protein-ligand interactions to assess ligand binding strength, while the latter uses known ligands and assesses how similar a compound is to the known reference compound(s). But this classification doesn’t encompass the full granularity of virtual screening techniques, especially with the introduction of virtual enumerable libraries like Enamine’s REAL and the methods used to tackle these very large databases of billions of compounds.
In this article, we discuss some of the techniques available and demonstrate how, through the application of Cresset’s cutting-edge computational technologies, Cresset’s Discovery team can directly address individual project needs, rather than the generic application of a single method.
Virtual Screening Solutions
Virtual screening can be used to identify starting points for projects where no known starting point template is available. It is more commonly used to advance project development, typically via escaping from problematic chemotypes, identification of a backup series, IP protection or development of a novel IP position.
Early or late-stage toxicity issues, an incorrect physical property profile, selectivity issues, synthetic tractability, poor absorption, etc. can often become an impasse for a chemical series, especially if the problem arises within the core of the molecule, meaning that despite showing promising therapeutic application, a new chemical position is required. Even if a compound presents no obvious developmental problems, a chemically distinct backup series is always a practical approach. The protection of IP space is always a balance between number of compounds to exemplify, in order to secure the IP position, and the expense and time required. The opposite problem is what to do if your starting compound is within IP congested space, and how to move away from this region. Virtual screening can be used as a cost-effective solution to providing an escape into improved chemical space.
Virtual Screening Use Cases
At Cresset, our team of experienced computational chemists work with you to assess the project requirements and apply the most appropriate technique, or combination of techniques, to meet your objectives.
Fragment replacement screening is the replacement of a single region of a template molecule to address feature-specific issues. This can be undertaken using the reference ligand, protein structure or a combination of both against which to evaluate new molecules. This is typically done using a filtered, chemically relevant, library to facilitate synthetic expedience and to address the problems with the original functionality.
Structure-based virtual screening identifies compounds that are predicted to bind in the protein active site. This can identify new compounds using only the protein structure to evaluate their activity. This protein-based evaluation requires an identified active site to define the docking region. The Cresset Discovery team can carry out modelling of the biological target to fill in any gaps in the data and build a greater understanding of the system, often providing early insights which guide optimization further downstream. The results refinement is typically augmented using either chemical properties, ligand-based methods or both.
Ligand-based screening methods identify new compounds by comparing compounds to known active compounds, without explicit reference to the protein. This method is dependent on the correct identification of the template ligand(s) bioactive conformation(s). Post-processing typically includes structure-based evaluation, chemical properties filtration or a combination of both.
New modalities have been developed to feasibly scale 3D virtual screening methodologies to ultra large virtual libraries like the Enamine REAL collection, or similar. Working with libraries of this size means that traditional screening methods are too slow or expensive to be accessible even with the availability of low-cost cloud-based computing. These methods require additional post-filtering since, even at 0.1 - 1% compound recovery, as this still results in millions, or tens of millions of compounds to process. Typically, these results require additional filtration using structure-based, ligand-based methods or chemical properties to reduce the final number of compounds down to a manageable number.
The protein ligand complex used in the images above is the 7br3 crystal structure reference Wei Yan, Lin Cheng, Wei Wang, Chao Wu, Xin Yang, Xiaozhe Du, Liang Ma, Shiqian Qi, Yuquan Wei, Zhiliang Lu, Shengyong Yang & Zhenhua Shao, 2020 Nature Communications Article Number:5287
How to Maximize Return of Investment from a Virtual Screen
By having an in-depth understanding of the target system, current roadblocks and the overall project objective, the virtual screening strategy can be focused towards directly addressing the project problems. Clearly defining the objectives and requirements is the principal factor in the determination of virtual screen direction, as well as the ultimate success in the fulfilment of the project requirements.
References
Flare™, Cresset®, Litlington, Cambridgeshire, UK; https://www.cresset-group.com/flare/; Cheeseright T., Mackey M., Rose S., Vinter, A.; Molecular Field Extrema as Descriptors of Biological Activity: Definition and Validation J. Chem. Inf. Model. 2006, 46 (2), 665-676;