Statistical characterization of the random errors in the radioimmunoassay dose--response variable.

Area of Science:

• Analytical Chemistry
• Biochemistry
• Immunology

Background:

• Radioimmunoassay (RIA) is a widely used technique for quantifying substances.
• Accurate error assessment is crucial for reliable RIA results.
• Understanding error distribution is key for appropriate data analysis.

Purpose of the Study:

• To develop methods for evaluating random errors in RIA dose-response variables.
• To determine the relationship between error magnitude and position on the dose-response curve.
• To improve the accuracy of RIA data analysis and quality control.

Main Methods:

• Analysis of standards and unknowns in radioimmunoassays for cAMP and cGMP in triplicate.
• Calculation of mean, standard deviation, and variance for response variables at each dose level.
• Pooling of results from multiple assays and application of curve smoothing for error analysis.

Main Results:

• Confirmed theoretically predicted non-uniformity of variance in RIA data.
• Demonstrated that error magnitude varies across the dose-response curve.
• Established a relationship between error variance and response variable.

Conclusions:

• Weighted least-squares curve-fitting programs are essential for accurate RIA data analysis.
• Computerized methods can serve as pre-processors for routine RIA analysis.
• Accurate error evaluation enhances the reliability and quality control of radioimmunoassays.