Bio-Rad Gel Doc™ EZ System - Manual (Page 89)

75
Regression Method Calculations
Each regression method will calculate a standard curve. Some of the methods
provide the formula for the standard curve. In this case, the molecular weight
can be calculated by:
x = relative front of the band of interest,
y = molecular weight of the band of interest.
Linear (semi-log): The linear equation is: y = a + bx, where a is the intercept
and b is the slope of the line.
Note: The linear equation is calculated on the log of the molecular weight
values.
The R-square value may be used to determine the overall quality of the linear
fit. A linear regression with an R-squared value of greater than 0.99 is
considered a very good fit. The primary advantage of this method is that it is
extremely simple. The primary disadvantage is that it will deliver wrong results
if the data is not very linear.
Point-to-Point (semi-log): No single equation is available for the point-to-
point method. The slope of each segment of the curve between data points is
calculated independently.
Note: The log of the molecular weight values is used to calculate the slope for
each segment of the curve.
Logistic: The Logistic-4PL equation is:
where:
x is the mobility,
y is the molecular weight,
a is the estimated molecular weight at infinity,
b is the slope of the tangent at midpoint,
c is the midpoint, and
d is the estimated molecular weight at zero mobility.
Since the curve generated by the logistic-4PL regression method represents a
perfectly shaped S, it might not fit the data very well in all cases.
Cubic Spline: Cubic spline curves are smooth curves that go through every
data point. The model is a cubic polynomial on each interval between data
points. In some cases, a spline curve can work well as a standard curve for
interpolation. However, because the curve is calculated individually for every
pair of points, it does not correspond to any single equation.