Bicontinuous cubic lipidic materials are increasingly used as crystallization media for in meso crystallization of membrane proteins (MPs). Varying the lipid architecture may assist with encapsulation of larger proteins and promote crystal growth. However, not all lipids are compatible with the components of typical crystallization screens, and compatibility must therefore be checked prior to crystallization trials. The method currently used, high-throughput small-angle X-ray scattering (HT SAXS), may be time-consuming and is costly in valuable MP. We have therefore employed a modeling approach using Bayesian regularized neural networks to accurately predict the complex phase behavior of lipid materials under the influence of the PACT crystallization screen and determine the lipid characteristics that allow a lipid to retain a cubic phase under the multiple components required during an in meso crystallization trial. This information will be used to select robust lipids for use in crystallization trials and may allow for the rational design of new lipids, specifically for in meso crystallization.