Fig. 1

Comparison of proteomic identifications from the full reference database or the transcriptome-informed reduced database searches. A Graphical representation of the two MS/MS search strategies compared here. MS/MS spectra were assigned searching either the reference human Ensembl protein database (full protein DB) or a subset of the reference database, generated based on transcript expression (reduced protein DB), using the Mascot software. PSMs were first validated using the Proline software with the following prefilters: (i) PSMs with a score difference < 0.1 were considered of equal score and assigned to the same rank (pretty rank), (ii) only a single best-scoring PSM was retained per query (single PSM per rank), and (iii) minimum peptide length ≥ 7 amino acids. PSMs were then filtered at the score cutoff estimated by target-decoy competition for 1% FDR control. B Size and overlap of the reference human Ensembl protein database (full protein DB) and the sample-specific reduced transcriptome-informed protein databases (reduced protein DB). C Number of spectra (left) or peptides (right) exclusively identified in the reduced database (“additional in reduced DB,” in blue) or exclusively identified in the full database (“lost in reduced DB,” in red) searches