Fig. 2

Iterative search and spectral clustering strategies used in database construction. a In iterative search strategy, raw files are searched against the original database separately, then for each result file the corresponding proteins are extracted to generate a specific database. All raw files are searched against the specific database once more, and the final database composed with proteins identified in the second search. b Firstly, a delegate spectra list is created by the clustering algorithm, then through searching these spectra against the original database, a peptide list is generated. The reduced protein database construction depends on this list. c Left: the total dataset contains 32 MS runs, 1,953,239 MS/MS spectra, and after clustering, only 19.5% were reserved as delegate; middle: 95% peptide identifications from iterative searching strategy were also found in database derived from clustering strategy; right: the common identifications in two strategies represent 90% of the total amount, iterative search and spectral clustering strategies only have 8 and 2% unique peptides, respectively