Multiplexing samples in mass spectrometry-based proteomics has long been accomplished by isotopologues of small molecules. However, it may still be underappreciated that multiplexing 1000s of samples with mass tags does not actually require 1000s of isotopes, or 1000s of synthesis steps to create. We show in the preprint, now published Open Access in the Journal of Proteome Research, that plex sizes in the hundreds, an order of magnitude greater than state-of-the-art, are achievable using molecules comparable in size to existing commercial tags, and that going beyond hundreds may require larger molecules
This repository contains the scripts used to generate figures for the article.
The Python scripts in analysis/ run in the following order, each one using the files written by the ones before it:
generate_valid_formulas.pysamples random atomic compositions with a fixed seed and writes the 10,000 chemically valid formulas invalid_formulas_10k.csv.pubchem_query_10k.pyreads that file and queries PubChem for each formula, writing the match counts and example structures inpubchem_10k_results.csv(checkpointed, so re-runs resume rather than re-query).max_plex_and_plot.pycomputes the maximum plex size at 6 mDa for every formula and adds it back topubchem_10k_results.csvas themax_plexcolumn; both figure scripts need this column, so it must be run before them.plex_engine.pyis an imported library of the isotopologue-offset and plex-size calculations rather than a script to run on its own.valid_formulas_offset_boxplots.pyreads both CSVs and produces the minimum differential mass defect boxplots, for all valid formulas and for the PubChem-confirmed subset.two_panel_figure.pyreads both CSVs plusunimod.xmland produces the 6 mDa reagent synthesizability figures; it additionally requires RDKit and Pillow to draw the inset molecular structure.
Miscellaneous information, including publicly-available presentations about the article are available through: parallelsq.org/tagdesign.