Figure 2.
Histogram of similarity scores and three estimated normal curves. Mixture of metabolite standards: histogram of 3286 best scores obtained from searching the library and three estimated normal curves (red, green, blue) and mixture distribution (cyon). T1(red) and T2(green) are two components of the mixture density fT and F1(blue) is fF. The three component normal mixture for similarity score is the mixture of T1, T2, and F1. 43 true positives out of 52 are assigned to at least one sample spectrum and average score is calculated for metabolites matched more than one sample spectra. 43 average scores for the 43 true positives are indicated by black circle; For another example, we selected two metabolites with CAS number 107715-91-3 (TP1) and 6066826 (TN1). They have 8 and 2 match pairs, respectively. 8 scores for TP1 are indicated by red asterisks and 2 scores for TN1 are indicated by blue asterisks. Bigger asterisks (red, blue) indicate average of the scores with the same color.