. Author manuscript; available in PMC: 2011 Oct 1.

Published in final edited form as: Proteomics. 2010 Oct;10(19):3515–3524. doi: 10.1002/pmic.200900727

Table 2.

The features used in the study in order of their significance in classifying phosphorylated MS/MS data according to random forest classifier variable importance test.

Feature	Decrease in acuracy
Mascot score, S	0.760
Percent of unidentified peaks, N_noID	0.759
Mean intensity of b-ions, I_bIons	0.650
Maximum m/z value, mz_max	0.555
Number of low peaks, n	0.490
Mean intensity of fragment neutral loss peaks, I_NL	0.485
Intensity balance, I_bal	0.477
m/z value of maximum intensity peak, mz_Imax	0.475
Percent of unidentified peak intensity explained by neutral losses, NoID_nL	0.446
Mean intensity of y-ions, I_yIons	0.434
Total intensity, I_tot	0.423
Standard deviation of intensities, σ	0.422
Number of neutral losses, N_NL	0.368