Click here to close
Hello! We notice that you are using Internet Explorer, which is not supported by Xenbase and may cause the site to display incorrectly.
We suggest using a current version of Chrome,
FireFox, or Safari.
Nat Commun
2023 Dec 15;141:8365. doi: 10.1038/s41467-023-44139-5.
Show Gene links
Show Anatomy links
Mapping protein states and interactions across the tree of life with co-fractionation mass spectrometry.
Skinnider MA
,
Akinlaja MO
,
Foster LJ
.
???displayArticle.abstract???
We present CFdb, a harmonized resource of interaction proteomics data from 411 co-fractionation mass spectrometry (CF-MS) datasets spanning 21,703 fractions. Meta-analysis of this resource charts protein abundance, phosphorylation, and interactions throughout the tree of life, including a reference map of the human interactome. We show how large-scale CF-MS data can enhance analyses of individual CF-MS datasets, and exemplify this strategy by mapping the honey bee interactome.
Fig. 1. A harmonized resource of CF-MS data charts protein abundance and interactions across the tree of life.a Phylogenetic tree showing the 32 species with CF-MS data included in CFdb. b Expansions to the scope and coverage of CF-MS data in CFdb (“version 2”), as compared to our original meta-analysis (“version 1”). Phosphosite quantifications are assigned exclusively to version 2 because CF-MS datasets were not searched for phosphopeptides in our original meta-analysis. c Cumulative distribution function showing the number of fractions in which each human protein was quantified. Inset pie chart shows the total proportion of human proteins detected in at least one CF-MS fraction. d Abundance of human proteins detected by CF-MS in the original meta-analysis or the updated resource, versus those never detected by CF-MS, based on consensus protein abundance estimates from the PaxDb database61 (for n = 8,248 proteins overlapping between CFdb and PaxDb). e Tissue specificity of human proteins detected by CF-MS in the original meta-analysis or the updated resource, versus those never detected by CF-MS (for n = 10,978 proteins overlapping between CFdb and the Human Protein Atlas). f Precision of the human interactome inferred by meta-analysis of CF-MS experiments in CFdb as compared to our original meta-analysis, for interaction networks of a given size. g Precision of the human interactome inferred by meta-analysis of CF-MS experiments in CFdb for interaction networks of a given size, as compared to six high-throughput screens of the human interactome using Y2H or AP-MS. h–k, Comparisons to previous interactome screens highlight the quality of the CFdb human interactome. h Functional coherence of interactome networks, as quantified by the AUC of protein function prediction in cross-validation. Text shows the median AUC. Vertical lines show the proportion of GO terms with AUC less than 0.5, equivalent to random chance. i Coexpression of interacting protein pairs across a large proteomic dataset. Text shows the median Pearson correlation. Vertical lines show the proportion of negatively correlated pairs88. j Colocalization of interacting protein pairs by subcellular proteomics. k Connectivity between genes associated with the same disease, as quantified by the AUC of disease gene prediction in cross-validation. Source data are provided as a Source Data file.
Fig. 2. CFdb prioritizes functional phosphosites and enhances analysis of CF-MS data from non-model organisms.a Histogram showing the number of phosphoproteins quantified in each CF-MS experiment. b Proportion of phosphoserine (pS), phosphothreonine (pT) and phosphotyrosine (pY) residues in a large-scale meta-analysis of the human phosphoproteome, as compared to all phosphosites detected by CF-MS or phosphosites from the curated PhosphoSitePlus database16. c Functional scores of human phosphosites that were or were not ever detected by CF-MS (n = 116,258 phosphosites). d Functional scores of human phosphosites, stratified by the number of CF-MS fractions in which each phosphosite was detected (n = 116,258 phosphosites). e Number of phosphosites prioritized based on detection in at least five fractions across major non-human species or taxonomic groups in CFdb. f Example of a frequently quantified phosphosite, pS21 of TPI1. Chromatograms show the intensity of pS21-containing phosphopeptides (green) or the parent protein (light gray). g Overview of understudied human proteins for which an interaction or prioritized phosphosite was detected in CFdb. h Precision of the honey bee CF-MS interactome, inferred with and without a data augmentation strategy that leverages data from hundreds of experiments in CFdb. i Separation of intra- and intra-complex interactions in interactome networks reconstructed for four prokaryotes by random forest classifiers trained in cross-validation on species-specific protein complexes (“within-species”) versus on 206 human and mouse CF-MS experiments (“human/mouse”), as quantified by the area under the ROC curve (AUROC). Source data are provided as a Source Data file.
