ADMET

While drug design often starts with identifying a target and finding a compound with activity against the target, it doesn't stop there. Successfully finding a candidate which can be effective in human trials requires considering many other factors:

• The compound must be absorbed by the body successfully, both into the bloodstream and then into cells (for intracellular targets).
• The compound must get to the right place in the body (distribution).
• The compound shouldn’t be metabolized before it can act or interfere with important metabolic processes.
• The compound should be excreted from the body at an appropriate rate, which impacts dosing schedules &c.
• And the compound needs to not cause toxicity in patients, the most common cause of drug failure in clinical trials.

These considerations are often lumped together as ADMET, or ADME when toxicity is considered separately.

Predictive ADMET Models Can Help Drug Designers

ADMET issues—particularly toxicity—are one of the leading reasons why drugs fail clinical trials. Experimental ADMET optimization is often not a focus early in lead optimization: many ADMET assays require animal studies and a non-negligible amount of synthetic material, making it time-consuming and expensive to see how structural modification will modify cardiac toxicity experimentally.

Accurate predictive models can allow ADMET to be considered when considering new molecules to synthesize, putting these important properties front-and-center in exploration of chemical space. Quick estimates of whether various moieties might cause toxicity or metabolic side effects will prevent medicinal chemists from introducing liabilities early in the drug-design process, saving time and money later in the preclinical pipeline.

Rowan's ADMET Prediction Workflow

Rowan's ADMET workflow is built atop a state-of-the-art predictive model, ADMET-AI from Kyle Swanson and co-workers at Stanford. ADMET-AI combines graph neural networks and cheminformatic descriptors from RDKit to produce best-in-class results on the popular Therapeutic Data Commons datasets, as well as some other useful datasets. Our own tests show that, while the model isn't flawless, it generates useful and actionable results in matched molecular-pair experiments looking at hERG toxicity, CYP inhibition, and membrane permeability. (Read more about this on our blog.)

Rowan's interface automatically highlights potential liabilities to make it easy to interpret ADMET results at a glance—in the above calculation, it's easy to see that the compound in question is predicted to display potential hERG-induced cardiac toxicity (matching experimental results). Our ADMET calculations run in just seconds from SMILES or 3D structures, making it simple and painless to integrate these predictions into your existing workflow.