Synthetic peptide MS/MS analysis

To validate the fragmentation pattern of peptides assigned to altProts in the BioPlex dataset, a set of 100 tryptic peptides from 72 different altProts encoded by transcripts of various biotypes were synthesized (> 50% purity; Biomatik, Ontario, Canada) and subjected to LC-MS/MS analysis.

Two injections were prepared: one containing all synthetic peptides and the other containing a selection of 16 peptides from the first run that were undetected or only resulted in spectra of poor quality. First, the powder was resuspended in a solution of 1% formic acid and 50% acetonitrile. Then, the suspension was diluted to 20 nM in 1% formic acid and 5% acetonitrile prior to injection for the shotgun method (with the same parameters as those in the “MS analysis of in-house affinity purifications” section below) or injection for paired reaction monitoring (PRM) method (as published in [36]). Briefly, peptides were loaded and separated onto a nano high performance liquid chromatography (nanoHPLC) system (Catalog No. Dionex Ultimate 3000, ThermoFisher Scientific, Mississauga, Canada) with a constant flow of 4 μl/min onto a trap column [Acclaim PepMap100 C18 column (0.3 mm id × 5 mm), Dionex Corporation, Sunnyvale, CA]. Peptides were then eluted off toward an analytical column heated to 40 °C [PepMap C18 nano column (75 μm × 25 cm)] with a linear gradient of 5%–45% of solvent B (80% acetonitrile with 0.1% formic acid) over a 42-min gradient at a constant flow (450 nl/min).

Peptides were analyzed on an OrbiTrap Q Exactive (ThermoFisher Scientific) spectrometer using the PRM method. An inclusion list containing the m/z values corresponding to the monoisotopic form of the peptides was generated. The collision energy was set at 28% and resolution for the MS/MS was set at 35,000 for 200,000 ions with maximum filling time of 110 ms with an isolation window of 2.0. Data acquisition was conducted with Xcalibur version 4.3.73.11.

PSM was conducted using SearchGUI (version 3.3.17) and PeptideShaker (version 1.16.42) against the Swiss-Prot library (October 1, 2020) of proteins concatenated with the sequences of the 72 altProts (20,431 sequences) with FDR controlled at 1%. PSMs of synthetic peptides were then compared with PSMs observed in the BioPlex dataset using a spectral correlation measure as described by Toprak and his colleagues [72]. An example spectrum comparison (generated with the Universal Spectrum Explorer [73]) as well as an overall summary is available in Figure S2. All synthetic and BioPlex spectral comparisons are provided in Table S2.