binding-kinetics - SKILL.md Agent Skill

name: binding-kinetics description: Use when analyzing or predicting drug-target binding kinetics: kon, koff, KD, residence time, SPR data fitting (Langmuir/two-state), ITC thermodynamics, tau-RAMD and funnel metadynamics for unbinding, or kinetic QSAR models.

Binding Kinetics

Purpose

Analyze, predict, and optimize drug-target binding kinetics: on-rates (kon), off-rates (koff), residence time (RT = 1/koff), thermodynamic signatures (ΔH/ΔS), and structure-kinetics relationships (SKR).

When to Use This Skill

Analyzing SPR sensorgrams (Biacore/Sierra)
Fitting ITC thermograms for ΔH/ΔS/ΔG
Computing residence time from MD simulations (τRAMD, metadynamics)
Building QSAR models for koff/kon
Interpreting kinetic selectivity vs equilibrium selectivity
Prioritizing compounds by residence time, not just KD

Reference Files

File	Content
`references/kinetics-theory.md`	kon/koff/KD/RT definitions, kinetic selectivity, two-state binding, conformational selection vs induced fit, thermodynamic signatures
`references/spr-analysis.md`	SPR sensorgrams, 1:1 Langmuir fitting, two-state model, Rmax/Rtheor, bulk correction, Biacore data parsing, Python fitting
`references/itc-analysis.md`	ITC thermogram integration, n/KD/ΔH/ΔS/ΔG fitting, SEDPHAT equivalents in Python, van't Hoff, enthalpy-entropy compensation
`references/residence-time-md.md`	τRAMD (random acceleration MD), funnel metadynamics, WTmetaD koff estimation, HTMD τRAMD Python, unbinding pathway analysis
`references/kinetic-qsar.md`	Structure-kinetics relationships (SKR), features for koff/kon models, kinetic maps (LE vs kinetic efficiency), koff cliff detection

Quick Routing

"Fit my SPR data" → spr-analysis.md

"Fit my ITC experiment" → itc-analysis.md

"Compute residence time from MD" → residence-time-md.md

"Build a model to predict koff" → kinetic-qsar.md

"Why does my drug work despite poor KD?" → kinetics-theory.md

Key Relationships

# Core kinetic relationships
KD = koff / kon                    # M (equilibrium dissociation constant)
pKD = -log10(KD)                   # analogous to pIC50
RT = 1 / koff                      # seconds (residence time)
t_half = ln(2) / koff              # seconds (half-life of complex)

# Thermodynamics
ΔG = RT_gas * ln(KD)               # kcal/mol (RT_gas = 0.592 at 298K)
ΔG = ΔH - T*ΔS                    # enthalpy-entropy decomposition

# Typical drug ranges
# kon:  10^4 – 10^7 M^-1 s^-1
# koff: 10^-5 – 10^-1 s^-1
# KD:   nM – µM
# RT:   10 s – 10^5 s (hours)

Integration with ALKYL Skills

Docking poses for kinetic path analysis: docking skill
MD trajectories for τRAMD: force-fields skill + MDAnalysis
SKR feature computation: chem_props.py, chem_analyze.py
MMPA for koff SAR: mmpa skill
Uncertainty in kinetic predictions: uncertainty-qsar skill