Calculators

Total simulated time = number of steps × timestep. Fill in any two and the third is worked out for you (highlighted). Change either of the two you filled in and it updates.

GROMACS works in ps: dt = 0.002 is a 2 fs timestep and nsteps is the step count. NAMD works in fs: timestep 2.0 is a 2 fs timestep and run takes the step count.

A reaction's standard free energy ΔG fixes where its equilibrium sits. The two are linked by ΔG = -RT ln K, where R is the gas constant and T the temperature: a more negative ΔG means a larger K and a reaction that lies further towards products.

For an acid dissociating (HA ⇌ A^- + H⁺), that same free energy is the deprotonation ΔG, and K is the acidity constant K_a. Taking -log₁₀ gives pK_a = ΔG / (RT ln 10). A larger pK_a is a weaker acid that holds onto its proton more tightly.

At a given pH, the Henderson–Hasselbalch relation sets how much of the species is protonated: fraction protonated = 1 / (1 + 10^{pH - pKa}). When pH equals the pK_a the species is exactly half protonated. A pH one unit below the pK_a leaves it ~91% protonated, one unit above ~9%.

Binding affinity is reported in several interchangeable ways. The dissociation constant K_d is the ligand concentration at which half the receptor is bound, so a smaller K_d means tighter binding. pK_d = -log₁₀ K_d flips that into a bigger-is-better number: 1 µM is pK_d 6, 1 nM is 9.

The free energy of binding follows from ΔG = RT ln K_d, so each tenfold improvement in K_d is worth about 5.7 kJ/mol (1.4 kcal/mol) at room temperature. The same arithmetic is used for K_i and, less rigorously, IC₅₀.

Raw affinity rewards big, greasy molecules. Efficiency metrics normalise it by size or lipophilicity so that a small fragment can be compared fairly with a lead. LE = -ΔG / N_heavy is the binding free energy per heavy atom. A common target is 0.3 kcal/mol per atom or better.

LLE = pK_d - logP asks how much affinity comes from something other than hydrophobicity. A value of 5 or more is a healthy lead. BEI and SEI are the same idea per unit molecular weight and per unit polar surface area.

A dose in mg/kg says nothing about whether the compound reaches its target at a useful concentration. In a one-compartment picture the peak plasma concentration is C = dose × F / V_d, where F is the fraction absorbed and V_d the volume of distribution per kg. A drug that partitions heavily into tissue has a large V_d and a low plasma level.

Only the unbound fraction f_u is free to bind the target. Comparing that free concentration with the IC₅₀ gives a crude occupancy of C_free / (C_free + IC₅₀). This ignores clearance and time, so treat it as a back-of-envelope check rather than a PK model.

At thermal equilibrium the probability of finding a system in state i is p_i ∝ exp(-E_i / RT), normalised over all states. Only energy differences matter, so the lowest state is taken as zero.

A state 1 kT (about 2.5 kJ/mol at room temperature) above the ground state is e times less populated. A gap of 5.7 kJ/mol is a factor of ten. Paste conformer energies in to see which ones actually contribute. The curve is the Boltzmann factor itself, so every bar sits on it.

A barrier height sets how often a reaction or conformational change happens. The Eyring equation, k = (k_BT / h) exp(-ΔG‡ / RT), comes from transition state theory and needs only the free energy of activation. The prefactor k_BT / h is about 6 × 10¹² s⁻¹ at room temperature.

The empirical Arrhenius form, k = A exp(-E_a / RT), swaps that for a fitted pre-exponential factor. Either way, every 5.7 kJ/mol added to the barrier slows the rate tenfold. The mean waiting time 1/k tells you whether an event is reachable in an MD run.