Fig. 1.

DIA workflow comparison. Comparing a five-NP workflow (red) to a 19-concatenated-into-9 high-pH fractionation of depleted plasma strategy (blue), a plasma depletion strategy (green), and neat plasma (purple). (A) Step-by-step comparison of five-NP, high-pH fractionation, depleted, and neat plasma workflows. (B) Median number of protein groups identified by each workflow. Error bars denote SDs of assay replicates. The top dash depicts the number of identified proteins in any of the samples, and the lower dash represents the number of identified proteins in three out of three assay replicates (defined as the complete features). For this comparison, samples belonging to the respective workflows were processed as independent Spectronaut runs. When processed together, the median numbers of protein groups were 1,615, 862, 461, and 375 for five-NP, high-pH, depleted, and neat plasma, respectively. (C) CV of median-normalized peptide intensities filtered for three out of three identifications across assay replicates. Median CV is depicted on each plot. (D) Dynamic range of identified proteins matched with normalized protein intensities from a plasma protein database (22). Protein groups were filtered for complete features. Median log₁₀ intensity of complete features is shown on each boxplot, and the outliers are removed. (E) Percent coverage of the plasma protein database in each workflow (Top) and relative coverage of plasma protein database by the five-NP method to high-pH fractionation (Bottom) over negative protein log₁₀ intensities. The 95% interval is shown in gray. Protein groups were filtered for complete features. (F) UpSet plot showing the protein group overlap between the five-NP and high-pH workflows. Protein groups are filtered for complete features. The workflows included in each bar are shown as colored labels. (G) Comparison of number of proteins with specified functional annotations covered exclusively with five-NP (red), exclusively with high-pH workflow (blue), or both (overlapped). Protein groups were filtered for complete features. The Venn diagrams are proportional to the number of protein groups. Workflows were processed together using Spectronaut for all analyses except for A, in which each workflow was processed separately. All proteins and peptides were conservatively filtered at 1% protein and peptide FDR.