Welcome to webSLE
webSLE is a lightweight application designed to display clinical and transcriptional data from the Dallas pediatric SLE cohort
Clinical Data
webSLE enables the tracking of longitudinal individual clinical profiles. Select a patient and let customizable charts display clinical and laboratory parameters such as SLEDAI, anti-dsDNA antibody titers, complete blood counts, treatment and nephritis class at each visit. Understand the global clinical history of each individual at a single glance.
Mixed Models
Linear mixed models were developed to identify the transcriptional correlates of disease activity, SLEDAI component distribution and nephritis classes at the cohort level. Along with the SAS code for each model, webSLE displays the expression profiles and annotated lists of transcripts significantly modulated for each comparison conducted by the models.
Personalized Immunomonitoring
We leveraged Weighted Gene Co-expression Network Analysis (WGCNA) to develop an individual longitudinal immunomonitoring pipeline. webSLE displays module/trait correlation matrices, module eigengene profiles and module content per individual.
Blood Modules
Visualize changes in blood fingerprints over time with our blood module framework to identify alterations in leukocyte activation and/or frequency and track the status of major immune networks over time.
WGCNA Modules
Transcripts
GO Annotations
Mean Connectivity & Scale Independence
Traits & Modules Dendrograms
Modules Connectivity Dendrogram
Modules Connectivity heatmap
Welcome to the WebSLE.com tutorials
WebSLE.com is a lightweight web application built in R/Shiny that displays the clinical and transcriptional data from the Dallas Pediatric SLE Cohort. The interface complements the manuscript Longitudinal Blood Transcriptomics Uncovers Immune Networks That Stratify Lupus Patients. It also aims to foster further analysis of this dataset by members of the scientific and biomedical community within the context of their specific research interests. WebSLE.com is organized in 4 sections, including Clinical Data, Mixed Models, WGCNA Runs and Blood Modules, that are accessible from the top navigation bar.
These tutorials will show you how to:
- analyze clinical traits changes over time (for clinically-oriented use case scenarios)
- analyze continuous trait distribution per sample groups (e.g. SLEDAI distribution by treatment category)
- analyze the profile and content of each module identified by WGCNA
- identify the clinical traits that best correlate with module eigengene, through an intuitive module/trait correlation heatmap
- identify the transcripts that best correlate with clinical traits
- identify major module hubs, through module membership quantification
- browse aggregated data that overlays clinical trait and module profiles in a single chart
- export various datasets as .csv or .pdf formats
Clinical Data
The clinical data section is subdivided into 3 sections:
- Summary: displays summary information about the dataset, including demographic distribution, clinical information such as disease activity and treatment distribution, and laboratory measurements.
- Individual: lets the user select a patient and displays an array of clinical and laboratory measurements. The charts are customizable.
- Distribution: displays the distribution of continuous variables such as the SLEDAI or anti-dsDNA antibody titers grouped by race, treatment or analytical set.
Mixed Models
This section displays the results from the 3 mixed models developed for this study.
- Disease activity: this model aimed to identify transcripts correlated with DA, while accounting for the influence of patient race and treatment. This model specifically tested fixed effects for race, DA, treatment, age, race/DA interaction and treatment/DA interaction.
- SLEDAI components: this model aimed to identify differences between SLEDAI component groups, while accounting for patient race/DA and treatment. It included SLEDAI components, race, treatment, age, SLEDAI components/race interaction and SLEDAI components/treatment interaction.
- Nephritis classes: this model aimed to identify differentially expressed transcripts between proliferative and membranous lupus nephritis in response to MMF treatment. It included race, age and nephritis class/treatment interaction.
Under each tab, the user can select any one of the pairwise estimates conducted. For example, the estimate DA3 vs. DA1 ‐ Up in DA3 returns all transcripts that are differentially expressed between the disease activity groups DA3 and DA1, and specifically overexpressed in DA3.
WGCNA Runs
Modules
This section lists all the modules extracted by WGCNA for the patient selected. Modules are uniquely identified by a color. Upon selection of a module of interest, WebSLE displays:- a line chart representing the module pattern over time, with the module eigengene highlighted in red.
- a searchable table containing the list of transcripts from that module. The list of transcripts can be copied or exported in various formats.
- dynamically generated bar charts representing the Gene Ontology enrichment for biological processes, cellular components and molecular functions in that module.
Module/Trait Correlations
This section displays the correlation matrix between WGCNA eigengenes and both clinical traits (marked by a black square) and Baylor blood module profiles. Correlations with blood modules, which have been thoroughly annotated, can be used to facilitate the functional interpretation of the WGCNA modules. Typically, the user will start by looking at this heatmap to identify the best WGCNA correlates of a specific clinical trait (e.g. SLEDAI, anti dsDNA antibody titers, etc…) and refer back to the Module tab to analyze the content of these modules. The matrix is also displayed as a searchable table below for export and downstream analysis of correlation with third-party software.
Gene Significance
This section displays a searchable table of correlations between individual transcripts and clinical traits analyzed. Typically, the user will use this feature for two main purposes:- Quantify module membership (MM) for a specific transcript, i.e. identify transcripts within WGCNA modules that best represent the global pattern (eigengene) of the module they belong to. This WGCNA feature enables the identification of major hubs within these networks.
- Identify the transcripts that best correlate with a continuous clinical trait of interest, though gene significance analysis. This feature is useful to identify specific biomarkers of a particular clinical trait, which may not belong to the WGCNA module that best correlate with that trait.
Aggregator
This section lets the user overlay WGCNA module profiles with a customizable number of clinical traits on a single line chart. This feature is useful for presentations and reports.
Charts
This section displays a collection of charts generated by WGCNA, including:- Mean connectivity and Scale independence plots. These are used to determine some of the parameters used during the WGCNA run.
- The trait dendrogram, which highlights relationships between clinical traits.
- Module eigengene dendrogram and connectivity heatmap. These are useful to understand WGCNA module interconnectivity and to identify larger groups of modules that display similar patterns.
Blood Modules
This section displays the Baylor blood module expression profile for each visit for a chosen patient. The color scale represents the percentage of transcripts from the module that display a normalized fold change of at least 1.5 fold (up in red, down in blue) and a raw data difference of at least 100 as compared to the median of healthy controls. This feature is useful to get a quick overview of the expression dynamics of hematopoietic networks for each individual over time.
The transcript content of each module is displayed under the Annotations tab, in a searchable table.
Click on the following link to download the batch-corrected normalized data for the SLE longitudinal study.
SLE_Longitudinal_972_eset.RData