Challenges in deriving high-confidence protein identifications from data gathered by a HUPO plasma proteome collaborative study

The Human Proteome Organization (HUPO) recently completed the first large-scale collaborative study to characterize the human serum and plasma proteomes. The study was carried out in different locations and used diverse methods and instruments to compare and integrate tandem mass spectrometry (MS/MS) data on aliquots of pooled serum and plasma from healthy subjects. Liquid chromatography (LC)-MS/MS data sets from 18 laboratories were matched to the International Protein Index database, and an initial integration exercise resulted in 9,504 proteins identified with one or more peptides, and 3,020 proteins identified with two or more peptides. This article uses a rigorous statistical approach to take into account the length of coding regions in genes, and multiple hypothesis-testing techniques. On this basis, we now present a reduced set of 889 proteins identified with a confidence level of at least 95%. We also discuss the importance of such an integrated analysis in providing an accurate representation of a proteome as well as the value such data sets contain for the high-confidence identification of protein matches to novel exons, some of which may be localized in alternatively spliced forms of known plasma proteins and some in previously nonannotated gene sequences.