Figure 5From: Improved machine learning method for analysis of gas phase chemistry of peptidesAmino acid distributions for the third position of the observed peptides. Frequencies of amino acids in the third position are plotted as described in Figure 3. For well-predicted ions, proline (P) is much more common than in the general data base; for under-predicted ions, it is highly less common. The displayed statistic (see Figure 3 label) is based on a transformation of an assumed Poisson count that tends to have a distribution that is standard normal; hence, departures that exceed three in magnitude (location of dashed lines) might be considered unexpected.Back to article page