Identifying direct contacts between protein complex subunits from their conditional dependence in proteomics datasets

Abstract

Determining the three dimensional arrangement of proteins in a complex is highly beneficial for uncovering mechanistic function and interpreting genetic variation in coding genes comprising protein complexes. There are several methods for determining co-complex interactions between proteins, among them co-fractionation / mass spectrometry (CF-MS), but it remains difficult to identify directly contacting subunits within a multi-protein complex. Correlation analysis of CF-MS profiles shows promise in detecting protein complexes as a whole but is limited in its ability to infer direct physical contacts among proteins in sub-complexes. To identify direct protein-protein contacts within human protein complexes we learn a sparse conditional dependency graph from approximately 3,000 CF-MS experiments on human cell lines. We show substantial performance gains in estimating direct interactions compared to correlation analysis on a benchmark of large protein complexes with solved three-dimensional structures. We demonstrate the method’s value in determining the three dimensional arrangement of proteins by making predictions for complexes without known structure (the exocyst and tRNA multi-synthetase complex) and by establishing evidence for the structural position of a recently discovered component of the core human EKC/KEOPS complex, GON7/C14ORF142, providing a more complete 3D model of the complex. Direct contact prediction provides easily calculable additional structural information for large-scale protein complex mapping studies and should be broadly applicable across organisms as more CF-MS datasets become available.

Microbial networks in SPRING - Semi-parametric rank-based correlation and partial correlation estimation for quantitative microbiome data

“…In summary, we believe that, as quantitative microbiome profiling will become increasingly available, the semi-parametric rank-based estimators for correlation and partial correlation estimation discussed here provide an important tool for reliable statistical analysis of quantitative microbiome data. While we have focused here on targeted amplicon-based sequencing datasets, our methodology is broadly applicable to other biological high-throughput data with large excess of zero counts, including quantitative metagenomics (Satinsky et al, 2013), single-cell RNA-Seq data (see Risso et al (2018) for a recent statistical analysis framework), and mass spectrometry proteomics data (Drew et al, 2017). Moreover, the concept of latent correlation employed in SPRING can naturally generalize to joint analysis of multiomics dataset when, on the same sample, several zero-inflated data types are measured in tandem.…”

Mapping Functional Protein Neighborhoods in the Mouse Brain

“…A characteristic of PPIs measured by co-fractionation is that it is not known which interactions are direct and which are transitive ( i.e ., as between subunits of the same complex that do not directly contact one another) (Drew et al, 2017a(Drew et al, , 2017b . Though reporting the hierarchical tree of interactions, rather than flat clusters, can help disentangle this issue, XL-MS can provide empirical evidence of direct interactions, and our interaction map determined from XL-MS, though much smaller, still significantly overlaps the co-fractionation map ( Supplemental Figure 1 ).…”

Classification of Single Particles from Human Cell Extract Reveals Distinct Structures

“…In an effort to understand why proteins form these machines, proteome-wide studies have been conducted to determine the composition of protein complexes (Drew et al, 2017a; Gavin et al, 2002; Havugimana et al, 2012; Hein et al, 2015; Ho et al, 2002; Huttlin et al, 2015, 2017; Kastritis et al, 2017; Kristensen et al, 2012; Krogan et al, 2006; Wan et al, 2015). Similar studies have identified direct contacts between protein complex subunits computationally (Drew et al, 2017b) or by cross-linking mass spectrometry (Leitner et al, 2016; Liu and Heck, 2015; Rappsilber et al, 2000), and while these studies provide insightful predictions on protein-protein interactions, they lack directly observable structural information that can inform us on function and subunit stoichiometry.…”