Ligand Efficiency Metrics — Catalog
P0 reference — verbatim formulas and recommended ranges from 2014-schneider-de-novo-molecular-design-book §6.4.1 (Mazanetz, Law, Whittaker, pp. ~165–167) and §1.4 (Schneider & Baringhaus, Eq. 1.8). The metric stack used to prioritize fragments for elaboration and to monitor MPO progress through hit-to-lead.
LE itself: cited from Hopkins, Groom, Alex, Drug Discov. Today 9 (2004) 430–431 [50 / 29 in book].
Metric definitions (verbatim formulas)
1. Ligand Efficiency (LE) — Eq. 1.8 / §6.4.1 base
LE = ΔG / N = −RT ln(K_D) / N (kcal/mol per heavy atom)
Schneider 2014 §5.3.3 alternative form using IC₅₀:
LE ≈ −0.59 (ln IC₅₀) / HAC (at 298 K)
where HAC = heavy-atom count (non-hydrogen atoms), R = 1.987 × 10⁻³ kcal/(mol·K), T = absolute temperature.
2. Fit Quality (FQ) — Reynolds et al. [226]
FQ = LE / LE_Scale
LE_Scale = −0.064 + 0.873 · exp(−0.026 · HAC)
— Schneider 2014 §6.4.1 (1).
3. Size-Independent Ligand Efficiency (SILE) — Nissink [227]
SILE = affinity / HAC^0.3
where affinity is pK_i or pIC_50.
— Schneider 2014 §6.4.1 (2).
4. Group Efficiency (GE) — Verlinde / Ferenczy [228]
GE = −ΔΔG_b / ΔHAC
The binding-free-energy contribution of a functional group A→B addition, per heavy atom added.
— Schneider 2014 §6.4.1 (3). Caveat: assumes group-additivity; see Fragment Additivity Assumption and Superadditivity.
5. Ligand Lipophilicity Efficiency (LLE / LiPE) — Leeson [2]
LLE = −pIC₅₀ (or pK_i) − cLogP (or LogD)
— Schneider 2014 §6.4.1 (4).
6. Ligand-Efficiency-Dependent Lipophilicity (LELP) — Keserű & Makara [229]
LELP = log P / LE
— Schneider 2014 §6.4.1 (5).
7. Ligand Lipophilicity Efficiency AT (LLE_AT) — Astex / Mortenson & Murray [230]
LLE_AT = (0.11 · ln(10) · RT · (Log P − Log(activity))) / HAC (kcal/mol per heavy atom)
where activity is K_D or IC_50.
— Schneider 2014 §6.4.1 (6). The Astex-derived size-aware LLE; recommended over plain LLE for fragments.
8. Kinetic Efficiency (KE) [233]
KE = t_½ / (0.693 · HAC)
where t_½ is target-residence-time half-life. Late-stage-only metric.
— Schneider 2014 §6.4.1.
Recommended ranges (verbatim from §6.4.1)
| Metric | Range | Interpretation |
|---|---|---|
| LE | ≥0.3 kcal/mol per HA | typical fragment-elaboration target (drug at MW 500 Da, IC₅₀ = 10 nM, ~38 HA → LE = 0.3) |
| LE | ≈ −1.5 kcal/mol per HA at HAC ≥ 15 | binding-energy contribution plateau (Kuntz max-affinity ceiling) |
| Per-atom optimization gain | 0.29 kcal/mol per non-H atom | mean affinity addition per added heavy atom during MPO [146, 222] |
| FQ | ≈ 1.0 | optimal binding |
| FQ | < 0.6 | poor binding |
| SILE | > 2.5 | optimization target |
| LLE / LiPE | 5–7+ | aim during MPO; ensures non-promiscuous final compound |
| LELP | < 16.5 | ”Lipinski-compliant” ceiling |
| LELP (lead range) | −10 to +10 | optimize toward 0 with improving potency |
Method-choice survey (from §6.4.1 closing)
“A survey conducted on the blog web site ‘Practical Fragments’ has highlighted LE followed by LLE as being the two most popular metrics.”
— Schneider 2014 §6.4.1 [231].
The Mazanetz lab’s recommendation: use LE for fragment selection; track LE + LLE during optimization. LLE_AT is the preferred lipophilicity-aware metric at the fragment stage because it incorporates HAC.
How to use in STRC
- h01 phase 3b fragment-pocket-fit scoring: record LE per fragment in the ranked output; gate at LE ≥ 0.3 kcal/mol per HA.
- h01 v4 fragment-grow: track GE for each added group; if a substituent’s GE < 0.2 kcal/mol per HA, the addition is wasted size — revert.
- h01 phase 4 docking pose ranking: convert Vina ΔG (kcal/mol) and ligand HAC to LE; use LE-rank as a parallel sort key alongside raw Vina score (Vina-score alone biases toward larger ligands per §6.3.9 [185]).
- h01 phase 7+ pharmaceutical filter: require LLE_AT > 0.3 kcal/mol per HA at the lead stage as a defense against the Leeson “MW + log P creep” attrition pattern (Leeson 2011 [83]).
- Citation pattern: “ligand-efficient fragment scoring per Schneider 2014 §6.4.1 (Mazanetz et al.); LE original definition Hopkins, Groom, Alex 2004 Drug Discov. Today.”
- Caveat (per §6.4.1 [233]): KE (kinetic efficiency) is currently “of limited application to fragments” — fragments have intrinsically fast on/off kinetics. Don’t apply at hit stage.
Connections
[part-of]pharmacochaperone[source]2014-schneider-de-novo-molecular-design-book[applies]index[see-also]Lipinski Rule of Fives vs Congreve Rule of Threes Reference Table[see-also]Fragment Additivity Assumption and Superadditivity[see-also]Recipe — Fragment Library Filtering Pipeline[see-also]Recipe — Fragment Optimization Linking Merging Growing