Percepta: Interpreting predicted aqueous solubility vs experimental variability

Description
Predicted aqueous solubility values (e.g., logS) come with inherent uncertainty due to limited and variable experimental data. Differences between prediction and experiment should be considered in the context of typical experimental variability.

Solution

• Run aqueous solubility prediction (e.g., logS) in Percepta for your compounds.
• When comparing to experimental data:
  • Consider that different experimental protocols (pH, equilibrium time, method) can yield different solubility values for the same compound.
• Look at the magnitude of deviation:
  • Small differences (e.g., < 0.5–1 log units) may be within the combined model and experimental uncertainty.
  • Larger deviations warrant investigation.
• For large discrepancies:
  • Check whether the experimental measurement is done at a defined pH and compare to predicted pH‑dependent solubility, if available.
  • Confirm that the exact same tautomeric or ionization form was measured.
• Use predictions:
  • As a guide to rank or prioritize compounds.
  • Not as a substitute for definitive solubility measurements in critical formulations.