Abstract Summary We present RokaiXplorer, an intuitive web tool designed to address the scarcity of user-friendly solutions for proteomics and phospho-proteomics data analysis and visualization. RokaiXplorer streamlines data processing, analysis, and visualization through an interactive online interface, making it accessible to researchers without specialized training in proteomics or data science. With its comprehensive suite of modules, RokaiXplorer facilitates phospho-proteomic analysis at the level of phosphosites, proteins, kinases, biological processes, and pathways. The tool offers functionalities such as data normalization, statistical testing, activity inference, pathway enrichment, subgroup analysis, automated report generation, and multiple visualizations, including volcano plots, bar plots, heat maps, and network views. As a unique feature, RokaiXplorer allows researchers to effortlessly deploy their own data browsers, enabling interactive sharing of research data and findings. Overall, RokaiXplorer fills an important gap in phospho-proteomic data analysis by providing the ability to comprehensively analyze data at multiple levels within a single application. Availability and implementation Access RokaiXplorer at: [33]http://explorer.rokai.io. 1 Introduction In cellular signaling, protein phosphorylation, a vital post-translational modification (PTM), plays a key role in regulating protein activity, structure, and function. Dysregulation of phosphorylation is associated with various diseases ([34]Neddens et al. 2018). Kinases, which regulate phosho-proteomic activity, have emerged as key therapeutic targets ([35]Zarrin et al. 2021). While advanced mass spectrometry-based technologies have allowed extensive phospho-proteomic data generation, the availability of user-friendly tools for analysis and integration with other proteomic data remains limited. Our previous efforts, KSEA-App ([36]Wiredja et al. 2017) for kinase activity inference and RoKAI-App ([37]Yılmaz et al. 2021), which employs functional networks to enhance kinase inference, were among the first few online tools for phosho-proteomic data analysis. Other online tools, such as GiaPronto ([38]Weiner et al. 2018), Phospho-Analyst ([39]Zhang et al. 2023), PhosR ([40]Kim et al. 2021), MAPPINGS ([41]Adderley et al. 2022), and SQuAPP ([42]Ergin et al. 2022), have been developed recently for the analysis of phosphorylation data and other PTMs. Builds on the existing tools and recent developments in phospho-proteomics, we introduce RokaiXplorer, a comprehensive framework that facilitates exploratory analyses of protein expression and phosphorylation data in an online, interactive environment. As compared to the state-of-the-art, RokaiXplorer offers multiple unique features: (i) Multilevel analyses within a single application. RokaiXplorer assesses dysregulation in at the level of specific phosphorylation sites and individual proteins, employing dedicated statistical tests for site versus protein level assessments. It infers kinase activities using the RoKAI algorithm, enhancing coverage and accuracy through functional networks. With pathway enrichment analysis, RokaiXplorer provides insights into functional implications of dysregulated phosphorylation or expression profiles. (ii) Integrative analyses of proteomic and phospho-proteomic data. (iii) A “Live sharing” functionality, enabling labs to share their data and analysis results with other researchers interactively, promoting collaborative efforts and broadening accessibility to their findings. In addition to these unique features, RokaiXplorer also supports missing data, employs optimized statistical methods for swift analyses, and enables subgroup specific inferences. It offers interactively adjustable statistical cutoffs, allowing tailored analyses. RokaiXplorer’s unique features include an inspection window for detailed sample-wise investigation, and a report generator for exporting results. To facilitate reproducibility, RokaiXplorer allows users to save configurations and versions, thereby enhancing the tool’s utility for the scientific community. 2 Description of the application Rokai Xplorer is implemented with Shiny framework and offers a range of functionalities at five levels: (i) phosphosite, (ii) phospho-protein, (iii) protein expression, (iv) kinase activity, and (v) pathway enrichment ([43]Fig. 1). It allows for the identification of significant dysregulation at each level and presents top findings through various visualizations, including volcano plots, heat maps, bar plots, tables, and network views. One of the distinguishing features of RokaiXplorer is its interactivity, which enables users to click on selected items in the visualizations to access an inspection window. This window provides comprehensive information about the selected items, including the source of evidence for dysregulation, quantifications, and raw data for all samples. In addition, this inspection window allows the user to navigate across different levels of analysis. For example, when the user wants to investigate a kinase of interest, they can explore the phosphorylation of its target sites as well as its own phosphorylation sites and protein expression. Figure 1. [44]Figure 1. [45]Open in a new tab The workflow and key idea of RokaiXplorer. 2.1 Getting started Getting started with RokaiXplorer is straightforward and user-friendly. The application provides an interactive tutorial that guides users through the initial steps, making it easy to familiarize themselves with the tool. To begin using RokaiXplorer, users only need two types of input data: quantification data and metadata. The quantification data is a .csv file that contains the phosphorylation levels of each phosphosite/peptide, with each row representing a specific site and the columns containing quantification values for multiple samples. The metadata file complements the quantification data by providing additional information about the samples, such as their grouping. The main group field, which is mandatory, specifies the case/control status of the samples, while optional additional groups can be utilized to focus the analysis on specific subgroups if desired. RokaiXplorer supports data from all proteomics quantification methods (e.g. label-free, SILAC, isobaric labeling). Additionally, RokaiXplorer supports the input of protein expression data, enhancing its versatility for comprehensive analyses. 2.2 Main features Rokai Xplorer offers a comprehensive suite of modules that enable researchers to perform various analyses on their datasets, in addition to several features that improve user experience. * Multi-level analyses in one application: + Site-level phosphorylation: Computes fold changes and employs statistical tests, such as t-tests, to assess dysregulation and identify phosphosites associated with specific conditions or diseases. + Protein-level phosphorylation: Computes phospho-proteomic changes at the protein level, computing mean log-fold changes and estimating variance based on a Satterthwaite ([46]Satterthwaite 1946) approximation. The main advantage of this approach is the decreased missingness at the protein-level, which facilitates making comparisons between different experiments. + Protein expression: Focuses on protein expression and utilizes the same statistical pipeline as phosphorylation analysis. It is performed optionally, requiring an additional input file. + Kinase activity inference: Infers kinase activities based on observed dysregulation patterns of phosphosites using the RoKAI algorithm ([47]Yılmaz et al. 2021), which makes use of a functional network to improve the accuracy and robustness of the inference. + Pathway enrichment analysis: Identifies over-represented biological processes, molecular functions and gene ontology (GO) terms to provide insights into the functional implications of dysregulated phospho-proteomic or proteomic profiles. The enrichment analysis is performed using a chi-squared test with Yate’s correction to ensure reliable and statistically rigorous results. In addition, to decrease redundancy in the results, an option is provided to filter out highly similar pathways using the Jaccard index. * Multiple views and visualizations: Volcano plots for a collective view, bar plots for displaying top candidates, heat maps for more granularity, box plots for comparing subgroups, tables for statistical details, and network views for the exploration of functional neighborhood. * Interactive online interface: Allows users to perform proteomic and phospho-proteomic statistical analyses through an intuitive online graphical user interface. * One-click tutorial: Includes an interactive step-by-step tutorial to help users get started with ease. * Missing data support: Takes into account and handles any missing values during statistical calculations without requiring any filtering or imputation technique in advance. * Improved statistical power: Offers moderated t-test ([48]Smyth 2004) to identify differential dysregulation, which employs an empirical Bayes method to shrink the sample variances toward a common value and to augment the degrees of freedom for individual variances, resulting in improved statistical power. * Efficient computation: Employs optimization techniques and sparse matrix operations in the implementation to support real-time analyses that complete in a matter of seconds for typical data matrix sizes (at the order of 10^4 × 10^2 for phosphosites × samples matrix). * Adjustable statistical cutoffs: Provides interactive options to set statistical significance thresholds based on P-values, fold changes, or false discovery rate with Benjamini–Hochberg procedure ([49]Benjamini and Hochberg 1995). * Subgroup analysis: Allows tailoring the analysis to specific samples based on user-defined groups. * Reference proteomes: Supports analysis of three different species, Human (Homo sapiens), Mouse (Mus musculus), and Rat (Rattus norvegicus). * Augmented networks: Integrates interaction data from different sources, including kinase-substrate associations (KSA) from PhosphoSitePlus ([50]Hornbeck et al. 2015) and Signor ([51]Licata et al. 2020), co-evolution and structural distance from PTM code ([52]Minguez et al. 2015), protein–protein interactions from STRING ([53]Szklarczyk et al. 2021), and GO annotations from GO Consortium ([54]Aleksander et al. 2023). In addition, KSA network for different species are cross-mapped to increase their coverage (e.g. interactions based on evidence from human studies are included in the networks for mouse). * Exploring subgroup differences: Allows a specialized analysis where the difference between fold changes of two specific subgroups are compared to identify any group-specific dysregulation. * Inspection window: Provides sample-wise inferences and allows subgroup comparison for detailed inspection of any inference, with bar/box plot visualizations and downloadable data. Additionally, it allows quick navigation between different related biological entities (phosphosites, proteins, kinases, pathways, and so on). The user can explore the evidence for a kinase’s inferred activity by clicking on the nodes in the kinase networks, and the Kinase and Pathway tabs provide detailed evidence on kinase and pathway enrichment. * Reproducible analyses: Offers a feature to save and restore a configuration file, allowing users to effortlessly reproduce their analysis at a later time. In addition, users can select specific data versions and analyze earlier iterations through the interface. Alternatively, previous versions of the application can be executed locally using a single R command, ensuring seamless replication of analyses. For example, after loading the shiny library, the command to restore and run version “v0.8.0” is as follows: [MATH: runGitHub(“< mi mathvariant="monospace">rokaixplorer”, “serhan−yilmaz”, ref=“v0.8.0”) :MATH] * Interactive data browser: Provides the necessary groundwork for users to create and deploy their own interactive data browsers with ease! 2.3 Interactive data browser: share your discoveries! To facilitate collaboration and data sharing, RokaiXplorer offers an additional feature that enables researchers to deploy their own interactive applications showcasing their data and analysis results. With this feature, the applications can be accessed online with the user data and settings already pre-loaded, allowing collaborators and other researchers to freely explore their data and customize their analysis. Deploying RokaiXplorer with pre-loaded input data is a straightforward process. Researchers can easily prepare and deploy their applications by following a few steps using the provided R scripts in the Github repository ([55]https://github.com/serhan-yilmaz/RokaiXplorer). These steps include installing R, RStudio, and Rtools (for Windows users), creating an RStudio project, downloading the RokaiXplorer source code, and installing the required R libraries. Once the setup is complete, researchers can run RokaiXplorer in deployment mode, customize the application for their specific data and configuration, and make modifications to the application’s title, descriptions, and about page. Additionally, RokaiXplorer allows users to export configuration files, enabling them to set desired analysis parameters for the online application and ensure reproducibility of results. Finally, researchers can deploy their application to shinyapps.io, a popular platform for hosting and sharing Shiny applications. By setting up a shinyapps.io account, connecting it to RStudio, and deploying the application, researchers can freely and effortlessly share their interactive RokaiXplorer application with others through a unique link, making their findings accessible to a wider audience. 3 Sample application–Alzheimer’s disease We utilized the capabilities of RokaiXplorer to analyze proteome and phospho-proteome data from a mouse hippocampus tissue study on Alzheimer’s disease (AD) ([56]Yilmaz et al. 2024). The dataset includes variables such as time (3, 6, and 9 months), sex (male and female), and genetic background (5XFAD versus wild type), which correspond to specific AD phenotypes such as Aβ42 plaque deposition, memory deficits, and neuronal loss. The goal of this study was to explore temporal and sex-linked variations in AD, focusing on biomarker discovery and identification of potential clinical targets. This study involved various analyses to understand the phosphoproteome changes in the hippocampus of 5XFAD mice during AD progression. These included statistical analyses to identify specific dysregulated phosphopeptides between the case (5XFAD) and control (wild-type) mice groups, comparison of phosphorylation patterns to protein expression levels, investigating regulatory mechanisms involved in phosphorylation events through kinase inference analysis, as well as an enrichment analysis to understand the biological pathways and networks impacted by the observed phosphoproteome changes. To facilitate the interpretation of the findings and promote free exploration of the data and results by other researchers, we utilized the RokaiXplorer application to develop the interactive tool AD-Xplorer. The findings are presented in the form of a live data browser with analysis capabilities, which can be accessed online at: [57]https://yilmazs.shinyapps.io/ADXplorer. Overall, we anticipate that RokaiXplorer will be an appreciated tool in the community to analyze phospho-proteomic data because of its simplicity and speed, enabling the analysis of data at different levels in one application. RokaiXplorer is available at: [58]http://explorer.rokai.io. Acknowledgements