Dose response assays are generally characterized by a nonlinear relationship between the response and the amount of drug/agonist/antagonist given. When plotted, this response will usually exhibit a sigmoidal curve that is best described using the 4 Parameter Logistic (4-PL) or 5 Parameter Logistic (5-PL) nonlinear regression model equation.
Dose Response Curve
The C parameter in the 4-PL model represents the concentration value for the inflection point of the curve. Depending on whether you have an inhibition or activation assay, the C parameter also represents the half maximal inhibitory concentration (IC50) or the half maximal effective concentration (EC50) respectively. Due to the symmetry around the inflection point, the 4PL makes it very convenient to determine the EC50/IC50 value.
The 5PL is a different story because of the asymmetry around the inflection point of the curve. The formula for calculating the EC50/IC50 from the 5PL still involves the C parameter but also a couple of others:
- E parameter – This is fifth or the “asymmetry factor” for the 5PL model equation. A value of 1 for this parameter represents perfect symmetry around the inflection point making it equal to the 4PL model equation.
- B parameter – This is the Hill Slope and it a measure of the rapidity of the curve’s transition between the asymptotes.
The equation for calculating the EC50/IC50 from the 5PL is as follows:
x = (C(2^(1/E) – 1))^(1/B)
Anatomy of the 4 or 5 Parameter Logistic Model
MasterPlex ReaderFit offers full support for the 4 Parameter Logistic (4-PL) and 5 Parameter Logistic (5-PL) model equation making it an ideal tool for automated fitting of dose response curves and calculating the EC50 or IC50 values.
Here is a Quick Guide for this procedure:
- To acquire MasterPlex ReaderFit, download and install the free 14-day trial.
- Importing the raw response values is a simple copy & paste OR you can use the file import wizard to import Excel, .csv, or text files.
- This is a sample dose response data set which contains 3 different sets of data from which we will generate 3 dose response curves (each outlined with a different color). Each curve set contains 8 data points with duplicates that are side-by-side.
Sample Dose Response Data
- Select 1 entire set of data points (including replicates) and click the blue Sample Curve button. Please note that you can hold down the CTRL button and left click non-contiguous wells.
Marking Data Sets
- The Auto Fill dialog will pop up. This feature is only helpful if you have a serial dilution for your drug concentrations. If not, you will have to enter them in manually.
Auto Fill Feature for Entering in Serial Dilution Concentrations
In this example, Drug ABC was serially diluted from top (20,000 pg/mL) to bottom (0.26 pg/mL) with a dilution factor of 5.
- Click Apply & View to preview the values in plate tab to confirm.
- Repeat steps 2-4 for any other data sets you may have. In this example, there are 2 other data sets.
For each cell, you will notice 2 numbers: the response value on the top and the drug dose on the bottom. - Click on the Fit Curves tab.
- Make sure the Use Best Fit feature checkbox is unchecked and select either the 4-PL or 5-PL model equation for the curve fit and then click Calculate.
Model Equation Selection and Calculation
- MasterPlex ReaderFit will then draw the dose response curve(s) according to the parameters you have chosen.
Dose Response Curves based on the 4 Parameter Logistic
- To view the EC50 or IC50 calculations, click on the View Results tab and press the Show Columns Selector button.
- Drag and drop the EC50 / IC50 column to the table header row.
- (Optional) You can click on the Merge Cells By Group button for a cleaner output. This function basically merges cells of replicate groups and just outputs a single result instead of repeating the same value for all replicates.
- (Optional) If you have more than 1 dose response curve, you may want to group them together by dragging and dropping the Replicate Group Name column header onto Analyte Name.
The end results will look something like this:
For more information or to download the free 14-day trial, please visit the MasterPlex ReaderFit product page.
You may also be interested in reading our blog post on:
- MasterPlex Webinar Series
- 5PL nonlinear regression model
- 4PL nonlinear regression model
- Tips for ELISA analysis
- The importance of weighting with the 4PL and 5PL
- Dummies Guide to Curve-Fitting for Quantitative Analysis
- Blog posts
Hi:
I need to make a rabbit serum titration (1/100, 1/200, 1/400, …………1/51200) by indirect ELISA and I don´t know how can I calculate it with your software. Could you help me?
Reply
aliu reply on May 10th, 2010 10:03 am:
Hi Carmen,
MasterPlex ReaderFit can certainly help you with this type of curve fitting and I will be more than happy to help you out.
Do you need to interpolate concentration values or do you simply want to plot raw fluorescence values vs. dilution factor?
Allen
Reply
Anonymous reply on May 10th, 2010 11:59 am:
Only Absorbance values vs dilution factor
Reply
aliu reply on May 10th, 2010 12:23 pm:
In this case, first import your raw absorbance values. Then highlight and select all wells pertaining to a single curve (including all replicates) and then press the blue “S” button in the top row of icons. After that, the Auto Fill dialog will pop up. For “Starting Value,” enter “51200″ and use “2″ as your dilution factor. Then choose how many replicates you have per sample and the dilution direction of your plate.
Click “Apply and View” to make sure the dilution factors are inputted corrected. After this, simply press the “Best Fit” button and you should have your curve.
If you would like me to do the analysis for you, you can send me your raw data file (aliu@miraibio.com) and we can continue this conversation offline.
I hope this information helps.
Allen
Anonymous reply on May 11th, 2010 2:23 pm:
Hi aliu:
When I import my absorbance values i don´t have any problem but when I have to generate my standard curve the values changes, for example 1.24 (DO 450nm) change to 24…..and another question how is the best formule to calculate a serum titre : endpoint titer or EC50…a lot of question aliu
Carmen
aliu reply on May 11th, 2010 6:18 pm:
Hi Carmen,
I would definitely like to take a look at your project file to see if I can see what is happening. If you can, please email that to me.
For EC50 or IC50, I would use the 4 Parameter Logistic/4PL nonlinear regression model. Once you generate the standard curves, you will be interested in the C parameter as that will be your EC50 or IC50 values depending on the type of assay you are doing.
I hope this helps.
Allen
Hi,
I just wanted to point out that the link from your ELISA tips post to may be very confusing. The 5PL is not as easy to interpret as the 4PL. The “C” parameter in a 5PL curve is not going to be at the EC50 value.
Dan
Reply
aliu reply on June 10th, 2010 3:42 pm:
Hi Dan,
You are absolutely right! I had actually forgotten to put my little blurb on calculating the EC50/IC50 from the 5PL. I had it in a different blog post but simply forgot to paste it over. It is taken care of now. Thanks for notifying me of this.
Allen
Reply
aliu reply on June 10th, 2010 3:49 pm:
Great blog by the way =)
Reply
Dear Allen
Indeed, I have a drug need to study the pharmacokinetic of it, and before this I need to calculate the dose response curve of such drug, so my question is how can I do it ? , if I am going to use the experimnetal animal such as Rabbit, rats , …etc
Reply
aliu reply on January 10th, 2011 11:24 am:
Hi Eltayeb,
The blog post above includes a “Quick Guide” that shows you how to calculate the EC50/IC50 of your assay. If I misunderstood your question, please give me some more details about your issue and I’ll be more than happy to help out.
Allen Liu
Reply
Hi ,
What if i had only 3 data point but there is 50% reduction can I used this software to do that? If yes, how should I do that? If not, which software should I use?
Reply
aliu reply on March 18th, 2011 7:07 pm:
Hi mint,
Although, it would be ideal to have more data points, the curve fit will depend on the quality of your data set. I can definitely give your data set a try though to see if it will work. Can you give me your raw data? If you do not feel comfortable replying in this blog post, feel free to email me your data. My email address is “aliu at miraibio dot com”.
Thanks,
Allen
Reply
Add A Comment