By Dr. Michelle Liu, Abbvie

  Crystallization processes in pharmaceutical development are designed to reliably produce a solid form with a Pareto-optimal profile of multiple properties including stability, solubility, and processability. Because API molecules can often crystallize in more than one unique form, i.e., different polymorphs, hydrates, and solvates, a thorough understanding of the solid form landscape is needed, for the neat API as well as for potential salt or cocrystal forms. The concerted effort of both experimental and computational science and methods is required to control and derisk the selected solid form. In recent years the importance of in silico crystallization development has grown, thanks to the advancement of computational methods combined with improved compute resources. In silico approaches can provide atomistic and energetic insights into the thermodynamics and kinetics of a compound’s solid form landscape and crystallization, via both physics-based and data-driven methods. In this presentation, we will discuss the application of these methods in the context of ABBV-113, an oral NLRX1 agonist investigated for treatment of ulcerative colitis, showing how both computational and experimental approaches were used to assess and mitigate risks in crystallization development.