Updated project metadata. When evaluating the quantitative performance of label-free proteomics method(s), samples with known compositions are necessary to define key parameters such as reproducibility, missing data levels, accuracy, precision, as well as sensitivity/specificity for biomarker discovery. Here, we provide a carefully-designed multi-species spike-in sample set for such purposes, which was prepared by spiking small, variable amounts of E. Coli (for mimicking significant protein changes, i.e., true positives) and yeast (for balancing the variable levels of E. coli proteins) proteins into a large, constant background of human proteins (representing unchanged proteins). The sample set encompasses a total of 25 LC-MS sample runs (5 E.coli-level groups, 5 LC-MS replicate runs in each group). The proportion of human proteins is 60% across all samples, and each group contains the following percentage of E. coli and yeast proteins : A: 5%/35%, B: 7.5%/32.5%, C: 10%/30%, D: 15%/25%, E: 20%/20%. These files were used to comprehensively evaluate the quantitative performances by ultra-high-resolution (UHR)-IonStar and SWATH-MS in Result 3.2 of “Ultra-High-Resolution IonStar Strategy Enhancing Accuracy and Precision of MS1-Based Proteomics and an Extensive Comparison with State-of-the-Art SWATH-MS in Large-Cohort Quantification (DOI: 10.1021/acs.analchem.0c05002)”.