Fit dose-response and kinetics curves

What this involves

Fitting assay curves correctly requires choosing models that respect experimental error distributions. For enzyme kinetics, initial velocities must be determined strictly from progress curves where less than 10% of substrate is consumed. The resulting rates are fitted directly to the hyperbolic Michaelis-Menten model; linearising transformations such as Lineweaver–Burk plots distort experimental errors and are retained solely for graphical display.

For dose-response data, the agent fits a four-parameter logistic equation against log concentration, estimating the IC50 and Hill slope. If an assay does not achieve defined plateaus within the tested concentration range, an IC50 cannot be extrapolated and is reported strictly as a bound. Calculating an inhibition constant (Ki) from an IC50 requires specifying the inhibition mechanism and the substrate concentration relative to Km, applying Cheng–Prusoff solely for competitive systems.

All calculations run through the platform's agent using Python libraries including lmfit, scipy, and statsmodels on its own environment. Replicate wells are checked for plate-reader saturation or drifting baselines, technical replicates are pooled into biological units, and every fitted parameter is delivered with standard errors or confidence intervals alongside regression diagnostics.

How it goes

1. 1Raw data inspection and replicate handlingInspect progress curves and replicate wells to identify swapped positions, baseline drift, or detector saturation before averaging technical replicates into biological n values.
2. 2Initial velocity calculationExtract initial rates (v0) from the linear phase of progress curves, ensuring that less than 10% of the substrate has been consumed.
3. 3Nonlinear curve fitting with lmfitFit hyperbolic Michaelis–Menten or four-parameter logistic models directly to data points using lmfit and scipy, avoiding linearising transformations.
4. 4Diagnostics and uncertainty estimationCompute residual plots, parameter standard errors, and confidence intervals, flagging any parameter that hits an optimisation bound or spans an order of magnitude.
5. 5Derived kinetic parameters and database checksDerive kcat using active enzyme concentration or compute Ki via mechanism-specific models, optionally checking measured bioactivities against ChEMBL or BindingDB.

Start it here

Ask in plain language. The agent works on a Linux machine of its own — it installs and runs what the job needs, keeps every file, and reads the results back with the numbers explained. Nothing to set up, and a long run continues after you close the page.

What to watch for

• Calculating initial velocities from progress curves where more than 10% of substrate has been consumed distorts rate estimates.
• Using Lineweaver–Burk plots to calculate Km and Vmax distorts experimental errors; fits must be performed on the hyperbola directly.
• A dose-response curve that does not establish clear plateaus within the measured range has no IC50, only an upper or lower bound.
• Applying the Cheng–Prusoff equation to derive Ki assumes competitive inhibition and requires an accurate [S]/Km ratio.
• Testing a substrate range that does not bracket Km (roughly 0.2 to 5 times Km) leaves Km poorly determined.
• Treating technical wells as independent biological replicates produces falsely narrow confidence intervals.

What it runs

Questions