such as Bayer and AstraZeneca, to bring these advanced computational methods in‐house to more broadly accelerate their own discovery efforts [15–17]. These successes utilizing advanced computational modeling to accelerate discovery are also now being realized by other parties using different computational strategies, such as Exscientia and Sumitomo Dainippon Pharma, which will be jointly advancing a serotonin 5‐HT1A receptor agonist into clinical trials for obsessive compulsive disorder [18].
The success of these smaller biotech companies using advanced computational modeling to accelerate their discovery efforts, and the growing investments of large pharmaceutical companies to build out similar capabilities, raises a couple important questions: what specific modeling technologies are being most used, and how are they delivering such value? In this chapter, we will review those technologies that have made the strongest contributions to these recent successes, as well as provide opinions regarding how these technologies and their deployment may be improved in the future.
2.2 Essential Techniques
Excitingly, over the last decade, advanced computational methods have broadly matured to the point where they now can be used to accelerate all stages of preclinical drug discovery, including (i) Target Validation and Feasibility Assessment, (ii) Hit Discovery, (iii) Hit‐to‐Lead and Lead Optimization, and (iv) Preclinical Development. In this section we will review which computational techniques, in our view, have become indispensable components of computationally driven drug discovery and their primary use cases.
2.2.1 Target Validation and Feasibility Assessment
The selection of a target for a drug discovery campaign is an important and highly consequential decision. Selection of an inappropriate target may doom an otherwise highly successful discovery project to failure in clinical trials, if connection of the target to the underlying disease biology is misassessed. Likewise, there are many targets where identification of drug‐like small molecule ligands remains an unsolved challenge, and if the protein target structure presents the project team with no binding pockets appropriate to bind drug‐like small molecules, alternative therapeutic modalities, such as biologics might be required. Advanced computational methods can assist this process by refining intermediate resolution structures to a higher apparent resolution. This allows a project team considering the target to better understand the fine details of the challenges likely to be posed by the protein earlier in the discovery project than would typically be possible [19–21]. Once these improved structures are in hand, they can be used to pursue computational solvent analysis of the protein surface to identify regions of unfavorable solvation that might be exploited to develop high‐affinity small molecule ligands [22–26]. Likewise, computational fragment soak methodologies can be used both to infer the druggability of various sites at the surface of the protein, as well as investigate if cryptic pocket binding opportunities might be possible to explore [27–31]. Such analyses may provide the project team a firmer foundation from which to judge if a drug‐like small‐molecule therapeutic modality is an appropriate choice for a given target, and should further allow the project team to advance from target tractability assessment to hit discovery with confidence the identification of hit‐like small‐molecule ligands should be tractable.
2.2.2 Hit Discovery
If a project team intends to pursue a drug‐like small‐molecule therapeutic modality for a particular protein target, then an important first goal of the project will be to identify small molecules with appreciable affinity and characterizable structure–activity relationships. A wide variety of experimental techniques have historically been used to identify such developable ligand matter, including high‐throughput screening, fragment screening, and DNA encoded library screening [32–35]. In each of these methods, a sizeable pool of small molecules, perhaps a few thousand molecules in a fragment‐based screen to potentially billions of molecules in a DNA‐encoded library screen, are tested for affinity for the target protein. The most potent of these initial hits will then typically be resynthesized and re‐assayed to confirm their activity using lower‐throughput and more reliable experimental techniques. Given the long timelines and significant expense associated with pursuing these experimental hit finding strategies, it is perhaps unsurprising that virtual screening, i.e. the use of computational modeling to similarly evaluate such a candidate pool of molecules for affinity, is gaining ever broader acceptance [36, 37]. These virtual screening techniques include:
1 Molecular docking technologies where each ligand is oriented into the binding site of a structurally characterized protein and its likelihood of binding to the protein with appreciable affinity is estimated by an analysis of the interactions it may form with the protein [38–41].
2 Shape‐based screening where the 3D shape of the ligand is compared with the shape of other known active ligands for the protein [42–45].
3 Ligand‐based methods where a machine learning algorithm using the 2D structures of known active molecules learns in the training of the model to identify other molecules likely to exhibit affinity based on this earlier provided information [46, 47].
It should be further mentioned that hybrid methods blurring the lines between these different approaches to virtual screening are under active development, and the above categories are meant more to orient the reader to this body of work, rather than to provide a rigid taxonomy of methods [48]. Once a virtual screen has been performed, much like a traditional experimental screen, putative active molecules will be purchased or synthesized and then assayed to confirm their activity using lower‐throughput and more reliable experimental techniques. A successful virtual screening campaign will provide a diverse selection of hit molecules exhibiting affinity that will provide good starting points for the discovery project to initiate hit‐to‐lead efforts.
Rather than solely present experimental screening and virtual screening to find hit molecules as competing techniques, we also wish to highlight that hybrid methods using both experimental techniques and the broad toolkit provided by modern computational methods are being investigated to unlock challenging targets. For example, multiple groups have been exploring if the success rates of experimental DNA‐encoded library screens can be enhanced through application of machine learning technologies [49]. Likewise, free energy calculations of fragment linking might be able increase the number of action hits identified in experimental fragment‐based screens [50]. We anticipate further development of such mixed computational/experimental approaches to hit finding to be a very productive future research direction.
It is also appropriate to mention here that if small molecule ligands for a given protein have already been reported, it is common practice within the pharmaceutical industry to discover novel hits for the protein by way of scaffold hopping, where the central core of the earlier reported chemical matter is replaced by some other appropriate core either to resolve some deficiency in the earlier chemical matter or simply to avoid earlier intellectual property. Although this approach to hit finding is conceptually quite appealing, failure rates are quite high since the vast majority of alternative central core choices usually greatly reduce or eliminate the potency of the small molecule [51]. Highly predictive alchemical free energy calculations have emerged as a reliable solution to this challenge, and also appear well suited to guide macrocyclization modifications, and other more exotic approaches to derivative hit discovery [52–55].
2.2.3 Hit‐to‐lead and Lead Optimization
Once developable hits have been identified, a central goal for the project team will be to synthesize and characterize sets of congeneric compounds with developable structure–activity relationships and demonstrated in vivo efficacy, i.e. a lead series. And, once a lead series is identified, the project team will work to further identify a development candidate molecule within that lead series manifesting the property profile required for that molecule to be advanced with confidence into preclinical
