Interactive Shiny App: Mapping Biological Influences on the Human Plasma Proteome Beyond the Genome
We integrated genomic with deep phenomic information to identify modifiable and non-modifiable factors associated with 4,775 plasma proteins from the SomaScan v4 platform in ~8,000 mostly healthy individuals. We systematically characterised the joint contribution of over 63 biological, genetic and technical influences on the human plasma proteome. The comprehensive phenomic information included information on diet, objective measure of physical activity, body composition, medication and blood-based biomarkers.
The application offers a comprehensive and interactive platform for exploring how various biological, genetic, and technical factors influence plasma protein levels, utilizing a dataset of human plasma proteome. It is designed to facilitate insights into the complex relationships between these influences by offering the following key features:
- Single Protein Visualization:
- Users can select individual proteins to visualize the proportion of variance in plasma protein levels explained by various biological, genetic, and techical factors.
- The variance explained is presented in a color-coded bar chart, with different factors categorized (e.g., Anthropometric, Lipids, Cardiovascular, Glycaemic, etc.).
- Customizable options include toggling technical variables, adjusting Y-axis limits, and exploring detailed protein-specific data in interactive tables.
- Comparative Protein Analysis:
- This feature allows users to compare the variance explained for two selected proteins across different influencing factors.
- The plot shows side-by-side comparisons, enabling an intuitive understanding of how the same factors impact different proteins.
- Users can adjust the Y-axis limits and choose whether to include technical variables in the comparison.
- Variable Influence on Protein Ranking:
- Users can select a specific variable of interest (e.g., age, BMI, cholesterol levels) to visualize the top proteins for which that variable explains the largest proportion of variance.
- The app provides a bar chart showing the contribution of the selected variable, color-coded by the most explanatory factor category (e.g., dietary, genetic, inflammatory).
Technical Overview:
- The application is built using R Shiny and integrates various R libraries such as dplyr, ggplot2, and DT for interactive tables.
- The dataset consists of multiple variables related to biological traits, genetic risk scores (GRS), and technical factors, mapped to their influence on the human plasma proteome.
- Linear mixed models are used to compute the proportion of variance explained, providing a comprehensive overview of the influences on protein levels.
This app serves as a powerful tool, complementary to the journal publication for researchers and analysts aiming to explore the biological underpinnings of plasma proteome variation, with the flexibility to focus on single proteins, compare two proteins, or analyse variable influence on multiple proteins.
Data access
By using these results in your research, you agree to cite our publication. Our legal notice, data protection statement and data usage agreement apply.
