Great Results Start with Great Standard Curves

By Meghan Rego

Meghan Rego August 15, 2024

What do a viral vector production facility, food allergy testing lab, and the grad student down the hall from you have in common? All of them rely on standard curves in their day-to-day work. Indeed, viral vector production facilities frequently use qPCR with a standard curve to titer their viruses; food allergy labs use standard curves with ELISAs to measure allergen concentrations in food; and your friend the grad student relies on a standard in their bicinchoninic acid assay (BCA) to normalize samples before running a western blot. Standard curves are critical for a number of popular scientific applications, and the quality of a standard curve can make or break an experiment. Here, we will provide an overview of how to create and use a standard curve and provide some general considerations for scientists planning to use them in their assays.

What is a standard curve?

Scientists use standard curves to determine the concentration of a target molecule in an unknown sample. To do this, you serially dilute a control sample of known concentration and measure a specific response as a function of the concentration. You then plot the concentration of the standards versus the response and use that information to extrapolate the concentration of an unknown sample. For example, a BCA assay uses a serial dilution of a control protein, incubated alongside an unknown sample with the BCA reagent. The absorbance of both is then measured on a spectrophotometer and concentration versus absorbance is plotted for the known concentrations (standard curve), generating an equation that allows you to determine the concentration of the unknown sample.

How do I make the standards for a standard curve?

You create a standard curve by serially diluting a known control sample, called a standard. The ideal standard curve has at least five dilutions, with each step of the series diluted by the same factor. For example, Figure 1 depicts a 2-fold dilution series. Each step of the series dilutes by 1:2 for a series that ranges from 1:2–1:32. The specific dilution series used will depend on the expected concentration of the unknown sample. The unknown should fall somewhere in the middle of the standard curve. If you do not know what concentration to expect, you may need to run the experiment a few times and optimize your standard curve accordingly.

Graphic showing a series of tubes. Tube 1 has control at 1mg/mL. An arrow goes from it to Tube 2 (1:2) labeled "Step 1, 1:2). This is repeated for another four tubes, with each step showing 1:2 dilution for a final dilution of 1:32.

Figure 1: A 1 mg/mL control is diluted in a series ranging from 1:2 to 1:32. Each individual step of the series is consistent at 1:2. When preparing a dilution series, use a new pipette tip for each step and mix the samples well by vortexing or inversion. Created with BioRender.com.

When choosing a standard, make sure that the sample is pure and free of any contaminants that could affect the measurements. You'll need to prepare the standard, standard dilutions, and unknown sample(s) in the same buffer if possible, or very similar buffers if not, since buffer components can affect the final readout.

When preparing a standard curve, change pipette tips and mix thoroughly by inversion or vortexing between each step of the series. To increase the accuracy of the curve, avoid pipetting small volumes (< 2 µL) or volumes too large for a standard micropipette (> 1,000 µL). The standard dilution series should be run in duplicate or triplicate. The closeness of the data points generated by replicate values provides useful information about the accuracy of the curve, which we will discuss later.

Standard curves must be included every time the assay is run, ideally with freshly prepared buffer. This is because standard curves vary depending on a number of factors, including the user, equipment, assay incubation time, etc., which can vary even when the same protocol is being followed.

Finally, standard curves must be validated for each run. To validate, include a positive control of known concentration in the experiment and use the standard curve to calculate its experimental value. Though it varies from assay to assay, typically the experimental value should be within 15% of the expected value.

Subscribe to the Addgene blog

How do I plot and use a standard curve?

To create a standard curve, you graph the measured response of the serially diluted standard as a function of concentration on a scatter plot. For the BCA example, the concentration of each serial dilution is plotted on the x-axis and its absorbance is plotted on the y-axis. With the help of graphing software, you then generate a trendline for the data. The appropriate trendline varies between assays and may be linear or nonlinear. BCA assays have linear trendlines while those for ELISAs and other enzymatic assays are sigmoidal. Once the trendline is calculated, you can extrapolate the concentration of the unknown sample based on its measured response. An example of this is shown in Figure 2.

A graph with Concentration (mg/mL) on the x-axis and OD560 on the y-axis. On it, several dots, a trendline, and an equation are plotted. See figure legend for details.

Figure 2: The concentration of a serially diluted standard is plotted against its OD560 measurements and the trendline y = 0.6561x + 0.019 calculated. The concentration of an unknown sample is determined by using the trendline to extrapolate its x value, 0.963 mg/mL, from its y value, 0.651 nm.

How “good” is my standard curve?

A poor standard curve leads to over or underestimation of an unknown sample’s concentration. To determine the quality of a standard curve scientists use graphing software to calculate the data’s coefficient of determination, R². R² is the square of the correlation between the actual values and the predicate values and measures how well the data points fit the trendline. R² ranges from 0 to 1 with 1 being a perfect fit and 0 indicating that there is no linear relationship between the observed and predicted values. This could be because there is no correlation or it could mean that the relationship is non-linear and you should use a different method of analysis. Different fields and assays have varying criteria for an acceptable R² but for many scientific applications users aim for an R² above 0.95.

Additional considerations

When running assays that rely on a standard curve, it is critical that the unknown sample’s concentration lies within the dynamic range of the curve. The dynamic range is the linear span between the lowest and highest concentrations that the curve can accurately measure.

Dynamic range depends on a number of factors, including the dilution series of the standard and the minimum and maximum detection limits of the instrument being used to measure the response. The dynamic range needs to be wide enough to cover all possible concentrations of the unknown samples; however, it should not be too broad. A broad dynamic range covers a wider span of concentrations but suffers from decreased specificity, especially as it approaches the minimum and maximum thresholds. If the measurements for the unknown sample fall outside of the dynamic range of the curve, then the assay is invalid. If the unknown sample’s measurement is too low, redesign the standard curve dilution series with lower dilution factors. If the unknown sample’s measurement is too high, start with a higher concentration of standard or make smaller dilutions.

As mentioned above, it is a good idea to run the standard dilution series in duplicate or triplicate. Replicates allows you to calculate the curve’s coefficient of variation, or %CV. The %CV of a sample is the standard deviation of each replicate measurement divided by the mean of all the replicates multiplied by 100. To calculate the %CV of the standard curve, one averages the individual %CVs for each point on the curve. ‌The smaller the %CV, the more accurate the curve. Though it varies based on the assay, the ideal %CV is less than 15%. You should also run replicates for all of the samples in the assay, as this will increase the accuracy of the data.

For assays involving sigmoidal curves, manufacturers often recommend using advanced graphing software designed to plot complex trendlines. If you don’t have access to this kind of software, don’t give up! Rather than plotting the entire data set, plot only the linear portion. This allows you to use a simple linear regression trendline, which may be good enough if your unknown lies within the linear portion of the graph. It does restrict the dynamic range of your standard curve, so if your sample does not fall within this range, it may be time to reconsider your graphing program.

Summary

Few things in the lab are more disheartening than seeing a potentially exciting result quashed by a bad standard curve. Take the following precautions to set yourself up for success!

Study the instruments and specific assays you will be using ahead of time to see what factors may impact the dynamic range of the assay. Try to minimize these.
Review previous literature or results from labmates to get a sense of the concentrations you might expect to see for the unknowns in your experiment. Carefully set up a standard curve dilution series that takes into account the range of concentrations possible.
Include both a standard curve and a positive control in every run.
When setting up your dilution series, use consistent dilution factors, change pipette tips regularly, and mix samples very well.
Use duplicates or triplicates to increase the accuracy of your experiment.
Finally, carefully analyze the data and use the trendline that's the best fit.

Good luck, and may your R² never dip below 0.95!