Frequently Asked Questions

General

What is Gradiance?

Gradiance is an AI-powered analysis platform for life science researchers. It automates dose-response curve fitting, standard curve quantification, and other common assay analysis workflows using natural language instructions.

Do I need to know how to code?

No! Gradiance is designed for bench scientists. Simply upload your data and describe what you want in plain English. The AI handles all the statistical analysis and curve fitting automatically.

What file formats are supported?

Gradiance accepts CSV (.csv) and Excel (.xlsx) files. Your data should have at least two columns: one for concentrations/standards and one for measured responses.

Is my data secure?

Yes. Your data is stored securely and is only accessible to your account. We use industry-standard encryption and follow best practices for data security. You can delete your runs at any time.

Analysis & Features

What types of analysis does Gradiance support?

Currently supported:

• Dose-Response Analysis: IC₅₀/EC₅₀ with 4PL and 5PL fitting
• Standard Curves: Quantification with linear and nonlinear fits
• qPCR: ΔΔCt relative quantification
• Enzyme Kinetics: Michaelis-Menten (Kₘ, Vₘₐₓ) fitting
• Cell Viability: Percent viability and cytotoxicity analysis

What is the difference between 4PL and 5PL?

4PL (4-Parameter Logistic) fits a symmetric sigmoidal curve and is the standard for most biological assays. 5PL (5-Parameter Logistic)adds an asymmetry parameter and should be used when your curve is noticeably asymmetric around the IC₅₀.

How do I know if my fit is good?

Gradiance automatically calculates R² (goodness of fit) and runs quality control checks. For publication-quality results, aim for R² > 0.95. The AI will flag any quality issues with warnings.

Can I exclude outliers?

Yes! You can ask the AI to exclude specific data points. For example: "Exclude the point at 100 µM" or "Remove outliers with Z-score > 2". The AI will refit the curve without those points.

Can I fix parameters in the curve fit?

Yes. You can constrain the Top (maximum response) or Bottom (minimum response) to specific values. For example: "Fix the bottom to 0" or "Constrain the top to 100%".

Subscription & Pricing

Is Gradiance free?

Gradiance offers a free tier for basic analysis with rate limits. Paid plans provide higher monthly capacity, batch processing, team workspaces, and priority support.

What happens if I cancel my subscription?

You retain access to paid features until the end of your billing period. If your account is terminated, Gradiance retains your data for 30 days to allow export unless you request earlier deletion or retention is required by law.

Can I export my data?

Yes. Every run can be exported as:

• High-resolution plots (PNG at 300 DPI, SVG, PDF report)
• Reproducible Python scripts for every analysis type
• CSV and Excel files with fitted parameters and underlying data

Technical Questions

What statistical methods does Gradiance use?

Gradiance uses industry-standard nonlinear regression algorithms (Levenberg-Marquardt) implemented in Python's SciPy library. All fitting methods are peer-reviewed and widely accepted in scientific publications.

How are confidence intervals calculated?

95% confidence intervals are calculated from the covariance matrix of the fitted parameters using standard error propagation methods.

Can I access Gradiance programmatically?

An API is planned for future releases. This will allow you to integrate Gradiance into automated workflows and custom pipelines.

Does Gradiance work offline?

No, Gradiance is a cloud-based platform and requires an internet connection. This allows us to provide powerful AI analysis without requiring you to install any software.

Troubleshooting

My curve fit looks wrong. What should I do?

1. Check that your concentration and response columns are correctly detected
2. Verify that concentrations are in the correct units (not log-transformed)
3. Look for outliers that might be skewing the fit
4. Try asking the AI to use a different model (4PL vs 5PL)
5. Check for QC warnings in the results panel

The AI didn't detect my columns correctly. How do I fix this?

You can explicitly tell the AI which columns to use: "Use the 'Conc_uM' column for concentration and 'OD450' for response."

I got a QC warning. Is my result invalid?

Not necessarily. QC warnings are flags that you should review the fit carefully. Common warnings (like IC₅₀ extrapolation) don't invalidate the result but indicate potential limitations. Review the warning message and decide if the fit meets your quality standards.