The image below is a sample image for the paper showing the information derived from a single primer pair.
This next set of images shows the log(p) traces for each of the 6 primers overlaid on top of each other, referenced properly to the underlying MMTV base positions. The first image shows the entire [0,1200bp] range, and the next four images show 200bp sub-sections with overlapping primers. Where things don't get too confusing, I think that it is fairly clear that the overlapping primers are indeed providing redundant information, although there is some horizontal jitter because the alignment to the DNA sequence is only approximate and is not perfect. In order to combine all 6 traces to get maximum coverage, we will rank the 6 primers and where there is overlap, choose the result provided by the highest ranking primer. I tried two different ways of "scoring" these primers, and both yielded nearly the same end result. The first just looked at the mean log(p) across the region covered by that primer. The primer with the highest mean was ranked as the "best". The second method computed both the mean and the sigma, and then ranked the primers according to which had the max value for (mean-sigma).
These next set of plots show the four B2 traces obtained from three different primers that "should" be covering the same region of DNA. Two of the primers are "Left" primers and one is a "Right" primer. The traces from the two "Left" primers are clearly highly similar although there is some slight 2-5bp offset. The trace from the "Right" primer is harder to be sure that it is showing the same structure. Perhaps we would need to look at both the N2 and the B2 traces and verify that the same relative behavior is exhibited. But that we should be able to see by looking at the log(intensity-ratio) plots for the different primers.
This plot shows the smoothed log intensity-ratio curve, resulting from combining the information from all 6 primers, and then smoothing with a Gaussian smoothing function with window length N=41, 51, or 61. (The standard deviation of the Gaussian is equal to N/5, so sigma=8, 10, 12.) Those are the blue curves. The grey curves are the 6 different traces from the micro-array experiment (3 forward and 3 reverse). The units on the y-axis are the same : log2(intensity-ratio). It is rather remarkable, I think, that except for the peak that is at ~680 in the micro-array data seems to have moved over by 100 bases to ~580,the rest of the curves for the most part match remarkably well.