Most NMR software helps you analyse an experiment after you have run it. By then the expensive decision — whether to run it at all, and how — has already been made, usually from experience and a rough calculation on the back of something.
Our experiment planner exists because that calculation is doable properly, and because getting it wrong costs a week of spectrometer time.
Four verdicts
The planner computes plenty of numbers. But the numbers are not the point; the verdict is. Four of them come straight out of the kernel, and each one corresponds to an experiment that would otherwise have been run and thrown away.
Your KD is unfittable at these concentrations. If the protein concentration is well above the dissociation constant, the titration is stoichiometric: every ligand added binds, the curve is two straight lines, and the fit will return a KD with a confident error bar that means nothing. You will learn stoichiometry, which you probably already knew. The fix is to dilute, or to run a competition experiment — and it is much cheaper to hear that now.
Your peaks will vanish halfway through. Exchange is pseudo-first-order, so kex changes across a titration as free ligand accumulates. A titration routinely begins in intermediate exchange and ends in fast exchange, and the same amide can be fast in proton and intermediate in nitrogen simultaneously. The planner computes the regime at every point, for both nuclei, and tells you where peak height drops far enough that picking becomes unreliable.
Your sample is too dilute for the precision you want. Work backwards from the precision you need on a fitted rate, through the Cramér–Rao bound, to the signal-to-noise you need per point, and then to concentration and acquisition time. Five per cent on R2 needs a certain SNR; two per cent needs substantially more; one per cent needs more again, and the relationship is not linear. If your sample cannot reach it, no delay list will rescue you.
Your CPMG will exceed the probe's duty cycle. A long relaxation period at a high CPMG frequency is a lot of radiofrequency power. The planner checks the duty cycle against the hardware limit, because the alternative way to discover this involves an expensive probe.
Aimed at people who are not experts
The planner is deliberately built for non-expert users. It asks for the few things that genuinely bias the answer — protein concentration, oligomeric state — and takes order-of-magnitude guesses for the rest, because a KD estimate only needs to be roughly right to tell you whether the experiment is viable. Everything else has a declared default you can see and override.
Nothing is submitted or stored. The plan comes straight back.
_ end transmission //