Create a new solid-state option when the neat API, salt, or amorphous route cannot deliver the required product profile. CD ComputaBio combines AI-assisted coformer prioritization, supramolecular chemistry, crystal-energy analysis, and targeted experiments to design pharmaceutical cocrystals with a clear development purpose.
Avoid a broad screen when the API chemistry, product target, or competing salt pathway makes cocrystallization unlikely to add value.
Build a diverse, route-appropriate library that covers interaction motifs without wasting API on redundant candidates.
Move the most plausible API–coformer pairs to the front of the experimental queue while retaining chemically informative negatives.
Determine whether new API–coformer interactions can compete with strong self-association in the separate components.
Increase the chance of finding accessible phases by matching preparation routes to solubility and kinetic constraints.
Advance only cocrystals that show a relevant property gain, defensible identity, reproducible preparation, and acceptable phase stability.
A pair may cocrystallize but deliver no useful property change; another may look attractive computationally yet fail because the pure components outcompete the mixed lattice. We keep formation evidence, product value, and phase-control risk visible as separate decisions.
Translate the API limitation into measurable targets, then compare cocrystallization with salt, polymorph, amorphous, and formulation alternatives before committing material.
Integrate coformer safety, molecular-recognition motifs, hydrogen-bond propensity, electrostatic complementarity, ML ranking, and pure-component lattice stability into one prioritized set.
Challenge shortlisted pairs through mechanochemical and solution routes, multiple stoichiometries, and solvent conditions; confirm that any new phase is not a salt, solvate, eutectic, or physical mixture.
Compare target-property gain, dissociation risk, moisture and thermal stability, preparation reproducibility, coformer burden, and scale-up feasibility before nominating a primary and backup candidate.
The schematics below are original method-to-decision summaries created for this page; they do not reproduce publication figures or represent CD ComputaBio project results.
Sugden and colleagues combined computational and experimental coformer screening across 30 multicomponent systems. The study showed how computational elimination of low-probability pairs can conserve experimental effort while preserving discovery opportunities.[1]
View publicationWang and colleagues trained machine-learning models on cocrystal records and used the resulting ranking to guide a captopril coformer screen. Two experimentally prepared cocrystals were among the successfully predicted pairs.[2]
View publicationA useful screen is defined by the API's limitation, the target product profile, and the evidence needed to select a phase—not by the longest possible coformer list.
Cocrystals are particularly relevant when the API is nonionizable, a stable salt is unavailable, or salt formation does not solve the target property. For ionizable pairs near the salt–cocrystal continuum, proton-transfer evidence must be evaluated directly.
No. AI can rank API–coformer pairs and reduce experimental burden, but data bias, missing negative examples, solvent effects, stoichiometry, polymorphism, and kinetic accessibility require experimental validation.
The answer depends on molecular recognition features, allowed coformer space, API availability, preparation routes, and the property objective. A diverse, prioritized panel is usually more informative than a large redundant library.
We combine phase-specific PXRD with thermal and spectroscopic evidence, assess stoichiometry and proton-transfer indicators, and compare results with the separate components and prepared controls.
It must show a reproducible distinct phase, a meaningful advantage over the parent API, acceptable coformer context, scalable preparation potential, and manageable dissociation, conversion, moisture, and formulation risks.
Share your API structure, current solid-form limitations, product target, and available material. CD ComputaBio will propose a staged computational and experimental cocrystal design plan.
Submit your project details below, and our team will respond within 24 hours.
Talk to our technical team about your project!
I Want To Talk