LRE analysis of Real Time PCR data

LRE analysis (Rutledge R.G. BMC Biotech 2008) on raw data from a real time PCR machine

The MATLAB code reads the data from and Excel file. I wrote the code to read the output from our AB7500 system. Although data from other machines would work as long as the excel file is set up the same way. The format of the spreadsheet is: Col A - Well #, Col B - Detector, Col C - Sample name Col D and beyond - Cycle # with columns extending out to include however many cycles you are running and rows extending down to include however many wells you are using. The detector represents the primer set being used and the sample name is self explanatory (I have to input the sample names from another file). The values here are the fluorescence values minus the baseline (the DRn from the 7500 system software).
The cycle numbers need to be in text format not numbers alone. The well numbers need to be numbers, not A1 – H12. The Detector and Sample name columns need to be text.
This is only semi automated data processing. Each well needs to be examined to ensure the correct positioning of the LRE window. The LRE window should start in the linear portion of the amplification curve, much in the same area a threshold would be set for the standard curve or ΔΔCT method.The initial analysis by the MATLAB code requires that the r2 value for the LRE window be above 0.985. If this placement is not acceptable the requirement is lessened to 0.85, I can then look for the LRE window that is in the right area and gives me the highest r2. All of these calculations come from Rutledge and Stewart BMC Biotechnology 2008.
The Get Data File button opens a dialogue box to allow you to select you data file. The dialogue box shows .xls and xlsx files. Once you open the data file, the file name with path appears in the top text box and sample name appears in the text box directly below the file name. The well number appears on the LRE Plot.
The GUI provides three grafts which are described in the Rutledge references.
For all plots blue dots represent the full data set for a given well and the red circles represent the points within the LRE window.
LRE Plot - Ec (cycle Efficiency) vs. Fc (Cycle fluorescence) Ec (cycle Efficiency)
F0 Plot – The predicted F0 (fluorescence at time zero) vs. the Fc (Cycle fluorescence)
Amp Curve - Fc (Cycle fluorescence) vs. Cycle #
The panel in between the two upper plots gives the cycle number the LRE window starts on, the # of cycles in the LRE window, the r2 value for the line defined by the section of the LRE plot within the LRE window (red circles) and the F0 predicted by the parameters defined by the above mentioned line.
The panel in the lower right hand corner gives you the options for adjusting the LRE window.Checking the first box accepts the analysis and moves to the next sample.Checking the second box returns to the previous samples, allowing you to compare between replicates. Checking the third box resets the analysis to the original LRE window.Checking the fourth box eliminates the well from analysis setting all values for this well to NaN. This can be undone by checking the third box. The button group at the bottom of this panel is where you make actual changes to what cycle number the LRE window starts on. The first button decreases the LRE window start by 1 cycle
The second button decreases the LRE window start by 10 cycles The third button increases the LRE window start by 1 cycle.
Once an LRE window has been accepted for each of the wells, hitting the Write data to file button, creates a new worksheet in the excel file named “final” that contains; Well, Detector, Sample Name, Emax, Fmax, LRE Start, LRE End LRE rsqrd, and F0 mean.
The rest of the calculations are relatively simple to do in excel. For the Lambda controls, the M0 (ng of amplicon) is calculated as the[ (ng Lambda DNA/Rxn] * amp size(bp)] / bp total genome. The OCF is then calclulated as the F0 (from the LRE analysis) / M0.
For each gene, including the housekeeping gene, the well M0 is calculated as the F0 / (OCF*0.5)
the well N0 = (M0 * 9.1x1011 ) / Amplicon size For a given gene, the replicates are then averaged to give a sample copy number which is then normalized to the housekeeping gene copy number for that sample.
If you don’t have the Lambda information, there is no OCF. So the F0mean for each well is divided by the amplicon size in base pairs, and then averaged across replicates giving a relative expression value for each sample. This relative expression value for each gene of interest is then normalized to this relative expression value for the housekeeping gene resulting in a normalized relative expression value.