Table 6.

Peptide identification results for 376 mass spectra. The experiments measured the success rate of our algorithm under different conditions: various sequence database sizes (0.5 million, 5 million, and 50 million amino acids), different numbers of de novo paths (1,5,10), and three types of searches (without PTMs, a search that simultaneously considers 10 types of PTMs but allows at most one modified amino acid in the peptide, and a search that considers 10 PTMs but allows up to two modified amino acids). The results are shown in terms of: TP - true positives (correct identifications made by the algorithm), FP - false positives (erroneous peptide identifications made the algorithm), and FN - false negatives (instances in which the algorithm did not return any peptide identification).

Decoy DB Size	# De Novo Paths	No PTMs			10 PTMs / 1 allowed			10 PTMs / 2 allowed
		% TP	% FP	% FN	% TP	% FP	% FN	% TP	% FP	% FN
0.5 M	1	0.904	0	0.096	0.862	0	0.138	0.859	0	0.141
	5	0.973	0	0.027	0.960	0.003	0.037	0.960	0.003	0.037
	10	0.984	0	0.016	0.971	0.003	0.026	0.971	0.003	0.026
5M	1	0.904	0	0.096	0.857	0.005	0.138	0.854	0.005	0.141
	5	0.971	0.003	0.026	0.952	0.013	0.035	0.949	0.016	0.035
	10	0.981	0.003	0.016	0.960	0.013	0.026	0.955	0.019	0.026
50 M	1	0.888	0.019	0.093	0.862	0.045	0.093	0.851	0.045	0.104
	5	0.952	0.021	0.027	0.920	0.059	0.021	0.915	0.056	0.029
	10	0.963	0.021	0.016	0.920	0.059	0.021	0.920	0.059	0.021