The human Plasma Proteome Project pilot phase aims to analyze serum and plasma specimens to elucidate specimen characteristics by various proteomic techniques to ensure sufficient sample quality for the HUPO main phase. We used our proprietary peptidomics technologies to analyze the samples distributed by HUPO. Peptidomics summarizes technologies for visualization, quantitation, and identification of the low-molecular-weight proteome (<15 kDa), the "peptidome." We analyzed all four HUPO specimens (EDTA plasma, citrate plasma, heparin plasma, and serum) from African- and Asian-American donors and compared them to in-house collected Caucasian specimens. One main finding focuses on the most suitable method of plasma specimen collection. Gentle platelet removal from plasma samples is beneficial for improved specificity. Platelet contamination or activation of platelets by low temperature prior to their removal leads to distinct and multiple peptide signals in plasma samples. Two different specimen collection protocols for platelet-poor plasma are recommended. Further emphasis is placed on the differences between plasma and serum on a peptidomic level. A large number of peptides, many of them in rather high abundance, are only present in serum and not detectable in plasma. This ex vivo generation of multiple peptides hampers discovery efforts and is caused by a variety of factors: the release of platelet-derived peptides, other peptides derived from cellular components or the clot, enzymatic activities of coagulation cascades, and other proteases. We conclude that specimen collection is a crucial step for successful peptide biomarker discovery in human blood samples. For analysis of the low-molecular-weight proteome, we recommend the use of platelet-depleted EDTA or citrate plasma.