BIOT6002 Lecture 16 Data Handling - Part 4 PDF

Data Handling - Part 4 BIOT6002: Lecture 16 Lecturer: Dr. Caroline A. Browne Nota Bene Several types of calculators are forbidden for use in the exam hall. Please note these calculators in the file uploaded on Canvas: Permitted and Prohibited Calculators for Examinations. Measurements o...

Data Handling - Part 4 BIOT6002: Lecture 16 Lecturer: Dr. Caroline A. Browne Nota Bene Several types of calculators are forbidden for use in the exam hall. Please note these calculators in the file uploaded on Canvas: Permitted and Prohibited Calculators for Examinations. Measurements of Accuracy Calculate the accuracy of the following set of 8 replicates for the concentration of the protein of interest. True value 174 ng. 174.9, 140.8, 190.8, 204.9, 172.2, 210, 156.4, 162.6 Answer Error Rate = |Observed Value - Actual Value|/Actual Value × 100. 1.479 % Measurements of Accuracy Recovery is determined by spiking various sample matrices (typically serum, EDTA plasma, and heparin plasma) with a given amount of control analyte and measuring the average percent recovery. Generally, samples with expected recovery and linearity between 80-120% are acceptable. Matrix Recovery range (%) Average(%) Serum (n=5) 92-101 95 EDTA plasma (n=5) 89-104 98 Heparin plasma (n=5) 86-104 95 Recovery In a recovery experiment for an immunoassay, you are asked to calculate the % recovery for the following : Two pairs of test samples are prepared: 1. Add an equal amount of sample matrix to both tubes 2. Spike Tube 1 with analyte and tube 2 with water (equal volumes)) 3. Concentration of Analyte added to tube 1 = 100 mg/ml 4. Measure both tubes for analyte in the immunoassay The following set of data is produced: Concentration of analyte measured in tube 1 = 110 mg/ml Concentration of analyte measured in tube 2 = 0 mg/ml % recovery = conc. tube 1 – conc. tube 2 x 100 conc. analyte added to tube 1 = 110 – 0 x 100 = 110% = % recovery 100 Measurements of Accuracy Linearity is determined by spiking various sample matrices (typically serum, EDTA plasma, and heparin plasma) with a given amount of control analyte, performing four two-fold serial dilutions of each, and then measuring the average percent recovery. Sample 1:2 1:4 1:8 1:16 Serum (n=5) 91-98% 82-96% 78-104% 83-92% EDTA plasma (n=5) 93-105% 89-101% 88-97% 80-93% Heparin plasma (n=5) 90-103% 96-105% 97-105% 96-105% Calculating linearity The evaluation is based on the equation of a line that defines the relationship between the bias and the reference values of the parts or samples. To calculate the line of best fit, use the equation: y = ax + b where: y = bias value, a = slope of the line, x = reference value, b = the y-intercept To calculate the slope, a: Measurement System Linearity - Type A Uncertainty Calculating Linearity, where: n = total number of measurements made To calculate the y-intercept, b: With values for a and b, we can complete the regression equation (y = ax + b); it gives us the line of best fit. The bias is the value of the sample measurement minus the reference measurement Linear curve fitting The most straightforward way to analyse your immunoassay data is to use a linear regression curve fit. This generally means plotting the concentration vs. the assay readout (OD for ELISA ) Then use the standard equation, y = ax + b The concentration is generally represented as x, the assay readout as y, with a referring to the slope and b referring to the y-intercept where x = 0 The aim is to find values for the slope (a) and y-intercept (b) that minimize the absolute distance from the data point to the curve, also known as the residual. Non-linear curve fitting: 4-Parameter Logistic (4PL) Immunoassay standard curves typically produce an S-shaped sigmoidal curve, and different mathematical modelling called logistic regression. This allows for curve fitting beyond the linear range of the curve. This is the logistic range and is most simply described by a 4PL curve. This type of modelling still uses the underlying concept of summing the square of the residuals, but instead of minimizing residuals for a straight line, the S-shaped curve is defined by the following parameters.