Ligand efficiency metrics: LE, LLE, and LELP explained

The easiest way to make a compound bind harder is to make it bigger and greasier. Add a lipophilic group, bury more surface in the pocket, and the IC50 drops. The problem is that this kind of potency is borrowed, not earned: the same molecular changes that buy raw affinity also wreck solubility, permeability, metabolic stability, and safety margins. Ligand efficiency metrics exist to stop medicinal chemists from congratulating themselves on potency they will pay for later.

Ligand efficiency (LE): potency per atom

Ligand efficiency normalizes binding free energy by molecular size. The standard definition, from Hopkins and colleagues, is the binding free energy divided by the number of non-hydrogen (heavy) atoms:

• LE = -1.37 x log(IC50 or Kd) / HAC, where HAC is the heavy-atom count and the 1.37 factor converts pIC50 to kcal/mol at room temperature. LE comes out in kcal/mol per heavy atom.

A nanomolar inhibitor with 50 heavy atoms and a 100 nanomolar fragment with 18 heavy atoms can have the same LE, which tells you the small fragment is using every atom as productively as the big lead. A common rule of thumb is that an LE around 0.3 kcal/mol per heavy atom or better is a healthy starting point for a drug-sized molecule. LE is the metric that made fragment-based drug discovery rigorous: it lets you compare a weak, tiny fragment to a potent, large lead on equal footing and decide which one is the better chemical starting point.

Lipophilic ligand efficiency (LLE or LipE): potency per unit of grease

LE controls for size; LLE controls for lipophilicity, which is the property most tightly coupled to downstream attrition. The definition is deliberately simple:

• LLE = pIC50 - logP (or logD). It is the potency you have achieved over and above what raw greasiness would buy you.

The interpretation is intuitive once you frame it as a competition. A molecule with LLE near zero binds its target about as well as it partitions into octanol, meaning its affinity is essentially just lipophilicity in disguise. A candidate with LLE of 6 has a roughly million-fold preference for its target over the oily phase, which means it is binding through specific, designed interactions rather than nonspecific stickiness. Most quality oral drug leads aim for LLE in the 5 to 7 range. Watching LLE rather than IC50 during optimization keeps a series from drifting into the high-logP territory where hERG, CYP inhibition, and poor solubility cluster.

LELP: catching the metric that LE and LLE can both miss

LE and LLE can each be gamed. LELP combines them to catch a series that looks fine on size but is quietly buying its efficiency with fat:

• LELP = logP / LE. It asks how much lipophilicity you are spending per unit of ligand efficiency. Values roughly between -10 and +10 are considered the desirable window; large positive LELP flags a compound that is efficient per atom only because it is heavily lipophilic.

The point of carrying three numbers is that no single metric is sufficient. LE alone rewards shrinking the molecule; LLE alone rewards de-greasing it; LELP ties the two together so that a chemist cannot improve one at the silent expense of the other. Used together they turn lead optimization from a potency chase into a balanced negotiation among affinity, size, and lipophilicity.

Reading efficiency off a docking run

Docking gives you a predicted binding score, and it is tempting to rank a virtual library on that score alone. That reproduces exactly the trap these metrics were built to avoid: docking scores, like measured affinities, tend to reward larger and greasier molecules that fill the pocket with more contacts. Converting a docking score into a size-normalized efficiency, and pairing it with the computed logP of each candidate, gives a far more honest ranking, the molecules that bind well for their size and polarity rather than the ones that simply sprawl across the pocket.

Try the docking yourself

Efficiency metrics are most useful when you can see size, lipophilicity, and predicted binding for the same molecule at once. Open Studio and dock a small panel of analogs against your target. Read the predicted binding score alongside the heavy-atom count and the logP from the ADMET panel, and rank your analogs by efficiency rather than raw score. The compound that wins on potency-per-atom is usually a better lead than the one that simply scores hardest.

Liganx is molecular docking online: free, browser-based, and set up to show binding scores next to the physicochemical properties you need to compute efficiency. If you want to try molecular docking and read ligand efficiency off the results without a local install, that is the fastest path.

Primary sources

• Hopkins AL, Groom CR, Alex A. Ligand efficiency: a useful metric for lead selection. Drug Discov Today 9, 430-431 (2004). doi:10.1016/S1359-6446(04)03069-7
• Leeson PD, Springthorpe B. The influence of drug-like concepts on decision-making in medicinal chemistry. Nat Rev Drug Discov 6, 881-890 (2007). doi:10.1038/nrd2445
• Hopkins AL, Keseru GM, Leeson PD, Rees DC, Reynolds CH. The role of ligand efficiency metrics in drug discovery. Nat Rev Drug Discov 13, 105-121 (2014). doi:10.1038/nrd4163