12 replies to this topic

[ElsaElsa]

Hello, I just used a Mouse IGM ELISA kit from Bethyl that has limited analysis help.
I don't have much experience with ELISAs, but I do understand that the calibration curves are sigmoidal and linear regression does not give a good r^2 value. In the past I used to use EIA kits from Cayman and they have templates that customers can download for quantitaiton.

I do not have the software to generate a 4-parameter logistic curve fit

Can anyone tell me if I can obtain a general ELISA quantitation template from somewhere?

Also, what curve fit do you use? I have comtemplated using 2-order polynomial but then I am not sure how to apply the equation in Excel to quantitate my unknowns because I am not sure how to use some of the terms, for example:

y = 2E-09x^3 - 4E-06X^2 +0.0024x + 0.01 (what is the value of E???)

Any help will be higly appreciated   !

[mdfenko]

ElsaElsa, on Jun 4 2009, 12:28 PM, said:

y = 2E-09x^3 - 4E-06X^2 +0.0024x + 0.01 (what is the value of E???)

"E" means "exponent". 2E-09 means 2*10^-09. this shorthand started because you couldn't subscript or superscript on computers in the olden days. you will also find it used by calculators (except for most of the graphical ones).

[scistudent]

We use R&D sys ELISA kits, so maybe its different but we just do a linear regression with the standard curve and the unknowns come from this, we haven't had a problem with the r squared values.

[SatanClaus]

Isn't the IgM analysis a sandwich ELISA? If so, you should probably just do a simple linear regression. Remove the points with a high absorbance that do not fall on a straight line and plot it. Four and five parameter logistic curves are to my knowledge only useful when using competitive ELISAs...

[aliu]

The Hitachi Software MiraiBio Group has a piece of software called MasterPlex ReaderFit that analyzes Elisa data with various model equations including the 4 parameter logistic and the 5 parameter logistic.  There is a free demo version and it may save you all a lot of time and headaches.

For bioassay analysis, I would recommend the 5 parameter logistic though because it is an asymmetric function as compared with the 4 parameter logistic.  Here is the model equation for the 5 parameter logistic:

F(x) = A + (D/(1+(X/C)^B)^E)

Here is a blog post that go into this in more detail:

Bioassay Analysis with the Five Parameter Logistic (5-PL) Non-Linear Regression Curve-Fitting Model

I hope this information helps.

[march]

Hi Elsaelsa,
did you solve your equation,as i also face the same problem can you share what you learnt especially as to what E means

[mdfenko]

march, on Aug 11 2009, 12:10 AM, said:

can you share what you learnt especially as to what E means

look at post #2

[cognitator]

Dear all, I also have the same challenge. Using exactly the same kit and I am now unable to determine my sample conc. Is there any freely available software for this?

[DRT]

cognitator, on Aug 15 2009, 05:50 AM, said:

Dear all, I also have the same challenge. Using exactly the same kit and I am now unable to determine my sample conc. Is there any freely available software for this?

If you are stuck, don’t be scared to go back to drawing your standard curve by hand on graph paper. It is unlikely that you will lose much accuracy. Then at least you will know if your attempts on a spreadsheet are giving you the correct answers.

[Tappu]

hi use the formula y=mx+c
and then apply the formula on the grapg then you will be able to get some values , apply these values for the given formula ,
ex      y=0.0783x+0.0645 9 u will be able to get this value when u plot the graph and add the trenline and the formula]
once this is been done take the values and then apply them
so, x= (y-c)/m woulf be the formula need to use to appply for your results, i have been using excel to plot my graphs  
good luck

[MyAssays]

This page has everything you need to do a 4PL without needing to download, install or configure any software:

http://www.myassays....meter-fit.assay

Just paste in your raw data.  There's also an Add-in for Excel available at http://www.myassays....cel-add-in.html if you need to do it from Excel.

[aliu]

Hitachi Software has just launched a free online version of our desktop ELISA analysis software with 4PL & 5PL curve fitting models and much more: http://www.ReaderFit.com

Just like MasterPlex ReaderFit, the core curve fitting algorithms are 8+ years in the making and also includes weighting algorithms which are a necessity when dealing with heteroscedastic immunoassay data.

Here is a sample of the results you can get plus there will be MUCH more on the way as far as statistics go. Stay tuned! =)

Allen Liu
Hitachi Solutions America, Ltd.
aliu [at] miraibio [dot] com

[Martin Fitzpatrick]

If you're doing a linear regression for your standards you can use the Excel template mentioned in this answer to automatically calculate the formula for the standard and use it to quantify your OD values. The link has some instructions on how to use it, but it's pretty self explanatory.