**1. Introduction**

In the last decades, there has been considerable investment into creating sustainable chemical technologies and reactions, specifically within the pharmaceutical industry, in order to maximize the efficiency of the traditionally lengthy and expensive process of drug development [1–3]. Computational advancements have been developed in parallel to such efforts, namely to unravel the intrinsic entanglement between the crystalline material and its solid-state properties, both for the possibility of exploiting such knowledge to engineer a pharmaceutical drug with the desired properties, and also to ensure its stability and reliability [3,4]. Given that the orientation and interactions between molecules within a crystal determine a material's physiochemical characteristics such as solubility, any alterations in this arrangemen<sup>t</sup> could have substantial implications, especially in a pharmaceutical context in relation to bioavailability and shelf life. Numerous publications illustrate examples of how different computational techniques are being applied to investigate and predict such features [5–9].

This case study revolves around ganciclovir (Figure 1), which is an acyclic guanine nucleoside analogue, that acts as a DNA synthesis inhibitor to various forms of herpes viruses, thereby reducing the rate of viral growth within the host [10]. In general, ganciclovir is formulated as a sodium salt and is administered through an intravenous route (Cytovene®) or also as an implant (Vitrasert®) [11]. It is also used as an ophthalmic gel which contains ganciclovir in its pure form, and a very small percentage of water (Virgan®) [11,12]. In spite of its importance as a BCS class III antiviral, and most especially as a treatment

against cytomegalovirus, ganciclovir's physicochemical properties have still not been optimized [13,14]. They include poor permeability and limited bioavailability, all of which hinder the drug's performance. Due to such properties, this antiviral is administered in frequent and high doses, exposing the patient to higher risks of toxicity.

**Figure 1.** Chemical structure of ganciclovir.

In view of all the above, the need for a complete study arose, entailing a detailed analysis at intramolecular, intermolecular and supramolecular levels using computational methods. This investigation is a comparative study between two anhydrous, enantiotropically related conformational polymorphs I (USP reference standard) and II (Figure 2), in an attempt to identify the factors which influence polymorph stability and assess the possibility of forming other polymorphs [15,16]. Published experimental data characterizing ganciclovir polymorphs are used to complement the observations extracted from the developed methodology [15,17,18], which incorporates a combination of different statistical and thermodynamic approaches, namely the full set of programs and interfaces available through the Cambridge Structural Database (*Mercury*, *Mogul*, *Isostar*) and *CrystalExplorer*. This approach can be applied not only to the API (Active Pharmaceutical Ingredient) discussed in this article but also to other crystalline materials.

**Figure 2.** The molecular structure of: (**a**) anhydrous form I (CCDC refcode: UGIVAI01, 50% probability level); (**b**) anhydrous form II (CCDC refcode: UGIVAI, 50% probability level) [19]. Hydrogen atoms were omitted for clarity. The different temperatures at which diffraction data were collected (form I: 100 K, form II: 293 K) had no effect on the occurrence of the two crystalline forms themselves.
