I'm reading Chapter 6 of Gentleman et al.'s book. For the quality assessment of the 8 calibration spectra, the "retain" parameter has values 0.118-0.178 (on Page 102). The interpretation of "retain" is "the number of high peaks in a single spectrum". My question is, given the sparsity of peaks in a spectrum, how could the proportion of high peaks as large as above 10 percent? Thanks! -- output of sessionInfo(): qualRes <- quality(M.r, peakfile, cutoff = 400) -- Sent via the guest posting facility at bioconductor.org.