Abstract
Calculable physicochemical descriptors are a useful guide to assist compound design in medicinal chemistry. It is well established that controlling size, lipophilicity, hydrogen bonding, flexibility and shape, guided by descriptors that approximate to these properties, can greatly increase the chances of successful drug discovery. Many therapeutic targets and new modalities are incompatible with the optimal ranges of these properties and thus there is much interest in approaches to find oral drug candidates outside of this space. These considerations have been a focus for a while and hence we analysed the physicochemical properties of oral drugs approved by the FDA from 2000 to 2022 to assess if such concepts had influenced the output of the drug-discovery community. Our findings show that it is possible to find drug molecules that lie outside of the optimal descriptor ranges and that large molecules in particular (molecular weight >500 Da) can be oral drugs. The analysis suggests that this is more likely if lipophilicity, hydrogen bonding and flexibility are controlled. Crude physicochemical descriptors are useful in that regard but more accurate and robust means of understanding substructural classes, shape and conformation are likely to be required to improve the chances of success in this space.