Fig. 2: Benchmark of DIP-MS versus other techniques for interactome analysis.
From: DIP-MS: ultra-deep interaction proteomics for the deconvolution of protein complexes
a, Venn diagram showing the overlap in identified PFDN2 interactors in SEC–MS, AP–MS (across 11 core subunits of PFD and PAQosome) and DIP-MS from the curated list of known PFDN2 interactors. b, Radarplot showing the performance of DIP-MS over AP–MS. c, Empirical cumulative distribution function plot for PFD–PFDL associated proteins identified in the PFDN2 DIP-MS experiment (average across replicates) and the SEC–MS from ref. 18. The x axis represents the log10 MS2 abundance integrated over a protein coelution profile in the SEC–MS dimension, defined here as a protein peak. The y axis represents the proportion of protein peaks at a specific abundance. d, Number of coelution peaks per PFD–PFDL associated proteins, identified in both, SEC–MS and DIP-MS. The solid black line represents the median for SEC–MS (1) and for DIP-MS (2) while the y axis represents the kernel density for the SEC–MS (blue density) and the DIP-MS (red density). Protein peak was defined as a signal having minimal width of three fractions and minimal height above background of 0.2 on a 0–1 signal scale. e, GED matrix between different techniques and databases. Color code represents the GED similarity from highly dissimilar (yellow, small circles) to isomorph graphs (dark purple, large circles). f, Barplot showing the number interactions recovered by different methods. Error bars represent standard deviation with n being the total number of combinations possible from the correspondent bait number. Different techniques are highlighted by different bar colors. Data are presented as mean values ± s.d.
