Bioinformatic Analysis of Temporal and Spatial Proteome Alternations During Infections

Abstract

Microbial pathogens have evolved numerous mechanisms to hijack host’s systems, thus causing disease. This is mediated by alterations in the combined host-pathogen proteome in time and space. Mass spectrometry-based proteomics approaches have been developed and tailored to map disease progression. The result is complex multidimensional data that pose numerous analytic challenges for downstream interpretation. However, a systematic review of approaches for the downstream analysis of such data has been lacking in the field. In this review, we detail the steps of a typical temporal and spatial analysis, including data pre-processing steps (i.e., quality control, data normalization, the imputation of missing values, and dimensionality reduction), different statistical and machine learning approaches, validation, interpretation, and the extraction of biological information from mass spectrometry data. We also discuss current best practices for these steps based on a collection of independent studies to guide users in selecting the most suitable strategies for their dataset and analysis objectives. Moreover, we also compiled the list of commonly used R software packages for each step of the analysis. These could be easily integrated into one’s analysis pipeline. Furthermore, we guide readers through various analysis steps by applying these workflows to mock and host-pathogen interaction data from public datasets. The workflows presented in this review will serve as an introduction for data analysis novices, while also helping established users update their data analysis pipelines. We conclude the review by discussing future directions and developments in temporal and spatial proteomics and data analysis approaches. Data analysis codes, prepared for this review are available from https://github.com/BabuLab-UofR/TempSpac, where guidelines and sample datasets are also offered for testing purposes.

Application of Machine Learning in Spatial Proteomics

“…Following data preprocessing, visualization of the full preprocessed data set in one figure is a crucial step in the analysis pipeline . Unsupervised dimensionality reduction and clustering are two commonly used visualization methods.…”

“…It has been used in several early studies to predict localization based on protein sequence . In MS-based and imaging-based spatial proteomic data preprocessing, the kNN is often used to impute missing values . In MS-based spatial proteomics, supervised ML algorithms are commonly employed to build classifiers that associate unannotated proteins to specific subcellular compartments, such as the SVM, kNN, RF and NB .…”

“…The major obstacles in integrating data from multiple modalities are the existence of batch effects and the distinct feature spaces of different omics data. , Integrating MS-based and imaging-based spatial proteomics approaches allows simultaneous analyses of multiple organelles at different time points, which can help understand disease processes. Advanced data analysis pipelines for such spatiotemporal analyses are desired to better integrate spatial and temporal maps of proteome changes during disease progression. , Since the occurrence of disease changes multiple interacting components of biological systems (such as RNA, proteins, lipids, and metabolites), multiomics integration incorporating spatial proteomics can give unprecedented insights into cell functionality. , Combining spatial proteomics and other omics data enables comprehensive studies of the heterogeneity in various diseases, and numerous related investigations have been conducted. ,− The development of analytical tools that provide workflows for multimodal data integration is currently a big challenge . Moreover, the lack of unified and systematic data analysis tools hampers the comparative analyses of changes in protein spatial patterns under different conditions .…”

“…To make sense of these spatial proteomic data, elaborate methodologies are needed to analyze the data and get insights into the underlying biological interpretation . Numerous computational methods have been applied to deal with the complexity of spatial proteomic data . Historically, predictions of protein subcellular localization used amino acid sequence and extracted features as inputs to train classifiers. , It has been reported that the inaccuracy of protein localization inferred from sequence or similarity had a detrimental effect on the identification of protein clusters .…”

Phospho-proteomics reveals that RSK signaling is required for proliferation of natural killer cells stimulated with IL-2 or IL-15

“…Because our analysis of AUC compared phospho-peptides by their magnitude of signaling but not their kinetics, we next sought to identify different temporal patterns of phospho-peptide signaling downstream of IL-2 and IL-15. Using the most up-and downregulated phospho-peptides (i.e., top 10% by absolute value of AUC, n=724), we performed fuzzy C-means clustering analysis (Kumar and Futschik, 2007), a soft clustering approach which identifies common signaling trajectories while allowing phospho-peptides to belong to multiple clusters if appropriate (Hu et al, 2015;Rahmatbakhsh et al, 2021;Yang et al, 2015). This analysis revealed six distinct clusters with different kinetic profiles (Fig.…”