The crystal structures of these pharmaceutical molecules are predictable – don’t take any chances!
You are worried that a competitor may patent a novel crystal polymorph of your best selling drug compound? You are concerned that the unforeseen appearance of a new crystal form with reduced solubility interrupts production after your product has finally reached the market? You fear that your amorphous formulation may suddenly convert into a previously unobserved crystal form? Or you simply would like to find other crystal polymorphs with more appropriate physico-chemical properties for industrial processing? Then you should talk to us!
Since 2002, AMS has been developing proprietary software for crystal structure prediction. In 2007, AMS’ computer program GRACE was the first to produce correct predictions for all 4 test compounds in the 4th blind test of crystal structure prediction organized by the University of Cambridge and hosted by the Cambridge Crystallographic Data Centre. And in 2015, GRACE achieved major advances, such as the first ever prediction of an hydrated Chloride salt, at the 6th CCDC Blind Test of Crystal Structure Prediction which have implications for pharmaceutical development.
For each of the molecules shown above, we can screen 88 – 95% of all relevant crystal packings and produce an accurate stability ranking in only a few weeks or months. Should we find a new crystal polymorph in the energy window delimited by the already known crystal forms, we can suggest experiments that will promote the growth of the new form and hinder the growth of the existing forms.
But information about how to make the new form may already be hidden in existing data from high throughput screening experiments. Comparing experimental X-ray powder diffraction patterns to those calculated from the predicted crystal structures, only a few previously unexplained diffraction peaks may be enough to identify a new crystal form.
Should the prediction reveal a new, more stable form, it may not always be possible to actually crystallize it. But from the lattice energy difference between the most stable experimental and predicted crystal structures, it is possible to work out the expected solubility difference. Should we predict a significantly less soluble form, then you may be well advised to prepare an alternative formulation in case your chosen one gets into trouble.