latest news


Welcome to the new version of the real time PCR website. Little information has changed, but the primer list has been updated!


setting the baseline

In the beginning of your real-time PCR run, the amount of fluorescence is very low.  At this point, the fluorescence is below the level of detection of the machine.  As the product accumulates, and there is increased fluorescence, the machine can begin to detect the sample.  In the graph below, the sample amplifies in a normal position, because it is a gene that is expressed at normal levels.  The little red tabs are set here ot cycles 3 and 14. This defines the baseline for the run. You need to tell the program where this baseline begins and ends to get the best data from your run.

The problem happens when you have samples that amplify really early in the PCR.  For example, with GAPDH, actin, or 18SrRNA.  

The Applied Biosystems program has a default setting for the baseline of cycles 6-15.  It does not change this automatically (it MAY in newer versions).  You need to reset the baseline to reflect your data, in the case that the default setting is not correct.  Here's an example of why you need to change the default regions.  The following is an amplification of GAPDH, with four different baselines chosen.  Note that 6-15 is the normal choice. 

In this graph, I shifted the baseline to the right.  Notice how there is a kink in the data at cycle 16.  This is an indication that you've set the wrong baseline.  Also notice how the lines have different slopes in the plateau portion of your graph. 

This one is 2-15.  There is still a kink.

This is 2-11.  Still a kink, but looking much better.

This one is placed correctly. Notice how the standard curve now looks much better.

In instances where the amplification occurs later, you must change the baseline again, but further to the right.  When the picture looks pretty, you are ready to analyze the data.

So what happens when you analyze the data when it looks like the first picture?  You lose sensitivity and are collecting bad data.  Here's a regression comparison of 48 data points from this experiment, comparing the 3rd and 4th graph.  We have a 93.4% correlation. 

When you compare the 1st and 4th graph, here's what happens.

The correlation coefficient drops significantly, showing you that you get different data if you set the positions of the baseline incorrectly!