The Proteomic Landscape of Genome-Wide Genetic Perturbations

The 5k Yeast Proteome Project

Christoph B. Messner, Vadim Demichev, Julia Muenzner, Simran Aulakh, Natalie Barthel, Annika Röhl, Lucía Herrera-Domínguez , Anna-Sophia Egger, Stephan Kamrad, Jing Hou, Guihong Tan, Oliver Lemke, Enrica Calvani, Lukasz Szyrwiel, Michael Mülleder, Kathryn S. Lilley, Charles Boone, Georg Kustatscher, Markus Ralser (2023)

SUMMARY. Functional genomic strategies have become fundamental for annotating gene function and regulatory networks. Here, we combined functional genomics with proteomics by quantifying protein abundances in a genome-scale knock-out library in Saccharomyces cerevisiae, using data-independent acquisition mass spectrometry. We find that global protein expression is driven by a complex interplay of i) general biological properties, including translation rate, protein turnover, the formation of protein complexes, growth rate and genome architecture, followed by ii) functional properties, such as the connectivity of a protein in genetic, metabolic, and physical interaction networks. Moreover, we show that functional proteomics complements current gene annotation strategies through assessment of proteome profile similarity, protein covariation, and reverse proteome profiling. Thus, our study reveals principles that govern protein expression and provides a genome-spanning resource for annotating the proteome.

MORE INFORMATION: Messner, CB, et al., The Proteomic Landscape of Genome-Wide Genetic Perturbations, Cell (2023).

ONLINE RESOURCE

Access individual profiles from the 5K Yeast Proteome Project

We measured proteomes for 4,500 single knock-out yeast strains using data-independent acquisition (DIA) and microflow-LC (see manuscript). The purpose of this app is to make the individual profiles accessible to researchers. Please find below a short tutorial. For more information, please read our manuscript for more information. Messner CB et al (2023), Cell

Tutorial:

Search protein tab:

• Select a protein of interest in the drop-down menu. By default only proteins quantified in at least 80% of knock-outs can be selected. Unchecking the box 'Restrict search to proteins measured in > 80% of KOs' will remove this filter and show PC for proteins identified in > 50% KOs (proteome profiles will be still restricted to proteins identified in > 80% KOs).
• Functional Associations: Shows results of a functional enrichment analysis. The enrichment analysis is an over-representation analysis based on KOs in which the selected protein was expressed differentially (RPP) or strongly linked proteins in the protein covariation network (highest-scoring 1% of associations in the network) (PC).
• Reverse Proteome Profile (RPP): Volcano plot showing the abundance changes of the selected protein across the 4,500 knock-outs. Significance was calculated using linear modelling and empirical Bayes. The user can indicate if imputed values should be shown by ticking the respective box. Further, the user can remove knock-outs that have a strong response (> 50 proteins differentially abundant), which can make the RPP more specific.
• Protein Covariation (PC) Map: UMAP showing the proteins grouped based on their covariation across the knock-outs. Subcellular locations can be highlighted by clicking the respective box.

To show more information on the proteins/KOs in the plots, the user can either hover over the points or use the 'Box Select' tool (output will be shown in a table). In order to unselect the proteins/KOs, double click somewhere in the plot. The user can further highlight certain proteins/KOs by using the respective input box below the plots.

Search knock-out tab:

• Select a knock-out of interest in the drop-down menu. By default only 'responsive' knock-outs are shown (knock-outs that have more than 15 (median) proteins differentially expressed).
• Functional Associations: Results of a functional enrichment analysis. The enrichment analysis is an over-representation analysis based on the differentially expressed proteins in the selected knock-out (PP) or strongly linked KOs in the profile similarity network (highest-scoring 1% of associations in the network) (PC).
• Proteome Profile (PP): Volcano plot: protein abundance changes of proteins in the selected knock-out. Significance was calculated using linear modelling and empirical Bayes. The user can indicate if imputed values should be shown by ticking the respective box. Further, the user can remove proteins that are associated with growth (|r| > 0.2; pearson correlation coefficients of growth rates with protein abundance changes across all KO strains).
• Profile Similarity (PS) Map: Knock-outs are grouped based on their protein profile. Subcellular locations can be highlighted by clicking the respective box.

To show more information on the proteins/KOs in the plots, the user can either hover over the points or use the 'Box Select' tool (output will be shown in a table). In order to unselect the proteins/KOs, double click somewhere in the plot. The user can further highlight certain proteins/KOs by using the respective input box below the plots.

Nomenclature:

Proteins are labelled with a 'p' after the genename and knock-outs with '-d' (e.g. Vps1p and Vps1-d for protein and knock-out, respsecitvely)

Note:

For RPP and PP, duplicated strains were averaged in the differential expression analysis to avoid that the same gene is counted more than once in the overrepresentation analysis. Therefore, some points in the volcano plot are not 'aligned' with the other points in the volcano plots. In the PS and PC each KO is shown.