Generation of a zebrafish SWATH-MS spectral library to quantify 10,000 proteins

Abstract

Sequential window acquisition of all theoretical mass spectra (SWATH-MS) requires a spectral library to extract quantitative measurements from the mass spectrometry data acquired in data-independent acquisition mode (DIA). Large combined spectral libraries containing SWATH assays have been generated for humans and several other organisms, but so far no publicly available library exists for measuring the proteome of zebrafish, a rapidly emerging model system in biomedical research. Here, we present a large zebrafish SWATH spectral library to measure the abundance of 104,185 proteotypic peptides from 10,405 proteins. The library includes proteins expressed in 9 different zebrafish tissues (brain, eye, heart, intestine, liver, muscle, ovary, spleen, and testis) and provides an important new resource to quantify 40% of the protein-coding zebrafish genes. We employ this resource to quantify the proteome across brain, muscle, and liver and characterize divergent expression levels of paralogous proteins in different tissues. Data are available via ProteomeXchange (PXD010876, PXD010869) and SWATHAtlas (PASS01237).

Figure 1. Workflow of creating and using the SWATH spectral library.

Samples were prepared using the pressure-cycling (PCT) workflow. The spectral library was built from fragment ion spectra generated by data-dependent acquisition mass spectrometry (DDA-MS) from fractionated and unfractionated peptide samples. The spectral library was then used to analyze samples from 3 different tissues using the OpenSWATH workflow.

Figure 2. Characteristics of the zebrafish SWATH spectral library.

(a) Number of true positive (blue) and target (orange) protein identifications at a given protein FDR. The applied MAYU protein-FDR cutoff is shown with a vertical dashed line. (b) Number of true positive (blue) and target (orange) peptide identifications at a given protein FDR. The applied MAYU protein FDR-cutoff of 0.01 is shown with a vertical dashed line. The dotted-dashed line indicates the peptide FDR threshold (0.01) that was applied to all peptides of proteins passing the protein cutoff. (c) Contribution of the individual organs to the number of identified proteins by proteotypic peptides. (d) Coverage of the proteome for SWATH spectral libraries of different species^,,. The coverage was calculated using the number of protein identifications with proteotypic peptides against the total number of proteins present in the sequence database, or the numbers from the cited publication were used (marked with an asterisk). (e) Barplot of the proteotypic peptides per protein in the human and the zebrafish library.

Figure 3. Quantification of the zebrafish proteome across brain, muscle and liver.

(a) Comparison of the SWATH assay coordinates to quantify the 10,115 peptides common to both the zebrafish and human SWATH spectral library. Plotted are the difference in retention time normalized to iRT peptides for the 11,816 precursor ions present in both libraries, how many of the 6 fragment ions per precursor (70,896 fragment ions in total) are identical, and the correlation of the relative intensities for the 11,463 precursor ions with at least 4 overlapping transitions. The horizontal lines in the violin plots depict the quartiles. (b) Overlap of proteins quantified across three different zebrafish tissues. (c) Coefficient of variation of the quantified signal for protein abundance across 6 fish. (d) Protein abundance of ohnologues across the three zebrafish tissues. The error bars represent standard deviation of six different fish and the difference in abundance was tested using an unpaired t-test for ohnologues quantified in the same tissue (^∗∗∗adj. p-value<0.001, ^∗∗adj. p-value<0.01, ^∗adj. p-value<0.05, n.s. adj. p-value>0.05).