This data descriptor introduces a curated dataset generated through a methodology that merges differential cell lysis, paired with swift extraction/digestion processes and mass spectrometry-based proteomics, aimed at the specific detection and identification of pathogens such as Staphylococcus aureus, Pseudomonas aeruginosa, and Candida albicans directly from whole-blood samples. Our protocol offers a rapid and direct diagnostic alternative, circumventing the traditional culture processes and thus facilitating the timely management of diseases like sepsis. We demonstrate the utility of this dataset by proposing a biomarker panel, derived from the proteomic profiles of the aforementioned pathogens, and applying our differential cell lysis protocol to blood samples from 8 sepsis patients. We achieved a sensitivity of 87.5% using parallel reaction monitoring (PRM), thereby providing diagnostics within a seven-hour timeframe without the need for microbial enrichment culture. Moreover, this dataset demonstrates high reproducibility and minimal outliers, solidifying its role as a reference for benchmarking and development of bioinformatic tools for peptide panel in identification for microbial detection, antimicrobial resistance, and epidemiological studies.