The PreAtlas database provides the comprehensive resource and interactive analysis platform for exploring the vast amount of bulk and single-cell transcriptomic data, as well as curated
knowledges specially pertaining to diverse premalignant diseases.
To date, PreAtlas version 1.0 contains the gene expression data from a total of 4,857 samples and 1,955,023 cells from high-quality curated 86 bulk and 34 single-cell transcriptomic datasets, for 79 manually standardized premalignant disease entries for 19 cancer types across 15 body sites.
Via the web interface, uses can browse, search, perform online basic bioinformatics analysis upon and download manually curated normalized data and distilled knowledges.
To date, PreAtlas version 1.0 contains the gene expression data from a total of 4,857 samples and 1,955,023 cells from high-quality curated 86 bulk and 34 single-cell transcriptomic datasets, for 79 manually standardized premalignant disease entries for 19 cancer types across 15 body sites.
Via the web interface, uses can browse, search, perform online basic bioinformatics analysis upon and download manually curated normalized data and distilled knowledges.
Homepage
Once users visit PreAtlas website, logo and a brief introduction of PreAtlas are shown, along with a glimpse of data status, including curated premalignant disease across multiple organs of human body, statistics of curated transcriptome data related to them, recent news about the database and hyperlinks towards related database as data source or reference in integration.
Browse
Users can browse the entry terms of PreAtlas by clicking the “Browse” button on the navigation bar in any page within the site. By selecting any of the three buttons at the top ofthe page, tables of curated bulk (Bulk datasets), scRNA-seq transcriptome datasets (Single-cell datasets) and disease term entries (Curated knowledges) are rendered correspondinglywith limited attributes and links towards the source dataset or viewing details. Users can filter the browse dataset entries by applying checkboxes on the left panel in order toview entries related to premalignant lesions of certain interested organs only.
Moreover, users are able to obtain detailed information through clicking the view links in tables of bulk, scRNA-seq transcriptome datasets and knowledges of disease entries.
For each bulk dataset, in the detail page, we provide:
a) Pie chart of sample counts overview of premalignant lesions;
b) Volcano plot and downloadable table of differentially expressed genes of different stages versus normal control samples;
Notes: DEG: Differently Expressed Genes (Premalignant lesion vs Normal control;FDR < 0.05 & fold change > 1.5, student t-test). c) Dot plot heatmap and downloadable table of dynamic genes across premalignant lesions along the known tumorigenesis pathway;
Notes: Dynamic genes referred to genes whose expression show gradually increase (UP) or decrease (DOWN) trend along the evolution of premalignant diseases (FDR < 0.05, anova test). d) Dot plot heatmap and downloadable table of tumor microenvironment (TME) cell types deconvolution analysis.
Notes: The deconvolution analysis was performed by the CIBERSORTx platform (https://cibersortx.stanford.edu/). TME, tumor microenvironment. For each scRNA-seq dataset, in the detail page, we provide:
a) UMAP 2-D dimension reduction plot with both manually annotated cell clusters and disease lesions of all cells;
b) Violin plot and downloadable table of putative cell markers analysis;
c) Boxplot and downloadable table of cellular proportion analysis across disease lesions within the dataset;
d) Trajectory plot and downloadable table of pseudo-tumorigenesis trajectory analysis;
Notes: The pseudo-tumorigenesis trajectory analysis was performed by the Monocle platform (http://cole-trapnell-lab.github.io/monocle-release/) and the putative cellular dynamic genes referred to those genes whose expression showed significant gradual pattern along the putative pseudo-tumorigenesis trajectory (pseudotime) For each curated knowledge entry term of premalignant disease, we offer the disease ID mapped in various popular medical term database (HDO, HPO, MESH, ICD10CM, etc.), and solidreference for the very disease or lesion being a stage of premalignant lesion of certain cancer.
Moreover, users are able to obtain detailed information through clicking the view links in tables of bulk, scRNA-seq transcriptome datasets and knowledges of disease entries.
For each bulk dataset, in the detail page, we provide:
a) Pie chart of sample counts overview of premalignant lesions;
b) Volcano plot and downloadable table of differentially expressed genes of different stages versus normal control samples;
Notes: DEG: Differently Expressed Genes (Premalignant lesion vs Normal control;FDR < 0.05 & fold change > 1.5, student t-test). c) Dot plot heatmap and downloadable table of dynamic genes across premalignant lesions along the known tumorigenesis pathway;
Notes: Dynamic genes referred to genes whose expression show gradually increase (UP) or decrease (DOWN) trend along the evolution of premalignant diseases (FDR < 0.05, anova test). d) Dot plot heatmap and downloadable table of tumor microenvironment (TME) cell types deconvolution analysis.
Notes: The deconvolution analysis was performed by the CIBERSORTx platform (https://cibersortx.stanford.edu/). TME, tumor microenvironment. For each scRNA-seq dataset, in the detail page, we provide:
a) UMAP 2-D dimension reduction plot with both manually annotated cell clusters and disease lesions of all cells;
b) Violin plot and downloadable table of putative cell markers analysis;
c) Boxplot and downloadable table of cellular proportion analysis across disease lesions within the dataset;
d) Trajectory plot and downloadable table of pseudo-tumorigenesis trajectory analysis;
Notes: The pseudo-tumorigenesis trajectory analysis was performed by the Monocle platform (http://cole-trapnell-lab.github.io/monocle-release/) and the putative cellular dynamic genes referred to those genes whose expression showed significant gradual pattern along the putative pseudo-tumorigenesis trajectory (pseudotime) For each curated knowledge entry term of premalignant disease, we offer the disease ID mapped in various popular medical term database (HDO, HPO, MESH, ICD10CM, etc.), and solidreference for the very disease or lesion being a stage of premalignant lesion of certain cancer.
Search
By clicking the “Search” button on the navigation bar in any pages within PreAtlas. We provide three main search modules, including “Search by gene”, “Search by organ” and “Search by project”. Here, the “Search by gene” module is enabled to provide basic information, associated organs or premalignant diseases, associated cell types, associated drugs and network associations for any input gene symbol or Ensembl ID. The “Seach by organ” module is enable to provide curated related transcriptomic datasets, associated cellular components and their proportions, associated dynamic genes and associated therapeutic drugs for diverse premalignant disease of any input organ. The “Search by project” module provides precise investigation of database including transcriptome dataset, or all datasets of bulk or scRNA-seq.
Online analysis
Users can enter analysis page by clicking the “Analysis” button on the navigation bar on any page. Here, four interactive analytical and visualization modules haven been packaged in PreAtlas to in-depth investigate these datasets, including Gene Expression Analyze (GEA), Cellular Component Analyze (CCA), Dynamic Gene Analyze (DGA) and Multiple Network analyze (MNA). Here, the GEA module enable to dissect the gene-level distribution across diverse cell types and/or pathological lesions while the CCA Module enable to dissect the cellular-level distribution based on either cell type annotations within scRNA-seq data or deconvoluted TME cells within the bulk transcriptomic data. Of note, the DBA module could identify premalignant-related dynamic genes as potential cancer risk-screening and early-diagnosis biomarkers, which were defined as those showing gradually dysregulated (increase or decrease) expression patterns along the pathologically dynamic evolution or putative pseudo-tumorigenesis trajectories. Finally, the MNA module enable to perform the network analysis for systematically associating premalignant-related multiple information, including diseases, genes, cell types and drugs, allowing for uncovering the evolution and intervention mechanism underlying premalignant diseases from the holistic view.
Download
PreAtlas provides users with downloadable resources of curated knowledges of premalignant diseases, full results of DEG, DGA and putative cell markers for premalignant disease.