DROMA_DB is a comprehensive database creation and management system that converts diverse omics datasets into a structured SQLite database for drug response analysis. This project serves as the foundation for the DROMA ecosystem, providing efficient data storage and retrieval for multi-omics drug response studies across different model systems.

It is a part of DROMA project. Visit the official DROMA website for comprehensive documentation and interactive examples.

• 🗄️ Unified Database Structure: Converts heterogeneous Rda files into a structured SQLite database
• 📊 Project-Oriented Architecture: Organizes data by research projects for efficient access
• 🔬 Multi-Omics Support: Handles various molecular profile types (mRNA, CNV, mutations, methylation, proteomics)
• 💊 Drug Response Integration: Comprehensive treatment response data management
• 🏥 Model System Coverage: Supports Cell Lines, PDOs, PDXs, and PDCs
• ⚡ Optimized Queries: Indexed database for fast data retrieval
• 📋 Metadata Management: Maintains comprehensive sample and drug annotations

The DROMA database contains 21 projects with comprehensive omics and drug response data:

The table for details are in sql_db/DROMA_Projects_info_detail.csv

Ensure you have R (≥ 4.0.0) and the required packages:

# Install required dependencies
if (!requireNamespace("RSQLite", quietly = TRUE)) {
    install.packages("RSQLite")
}
if (!requireNamespace("DBI", quietly = TRUE)) {
    install.packages("DBI")
}

Download the DROMA data from Zenodo:

# Data is available at: https://doi.org/10.5281/zenodo.15055905
# Download and extract the data files to your desired directory

git clone https://github.com/mugpeng/DROMA_DB.git
cd DROMA_DB

# Load the DROMA_DB functions
source("function/function.R")

# Create SQLite database from Rda files
createDROMADatabase(
    db_path = "sql_db/droma.sqlite",
    rda_dir = "data",
    projects = NULL  # Include all projects
)

# Connect to the database
con <- connectDROMADatabase("sql_db/droma.sqlite")

# List all available projects
projects <- listDROMAProjects()
print(projects)

# List all database tables
tables <- listDROMADatabaseTables()
print(head(tables))

# Close connection when done
closeDROMADatabase()

Once the database is created, use it with the DROMA_Set package:

# Install DROMA_Set package
# devtools::install_github("mugpeng/DROMA_Set")
library(DROMA.Set)

# Create DromaSet objects from the database
gCSI <- createDromaSetFromDatabase("gCSI", "sql_db/droma.sqlite")
ccle <- createDromaSetFromDatabase("CCLE", "sql_db/droma.sqlite")

# Create MultiDromaSet for cross-project analysis
multi_set <- createMultiDromaSetFromDatabase(
    project_names = c("gCSI", "CCLE"),
    db_path = "sql_db/droma.sqlite"
)

Ensure your data directory contains:

• anno.Rda: Sample and drug annotations
• {datatype}.Rda: Data files (e.g., mRNA.Rda, drug.Rda, cnv.Rda)

# Create comprehensive database
createDROMADatabase(
    db_path = "path/to/droma.sqlite",
    rda_dir = "path/to/data",
    projects = c("CCLE", "gCSI", "GDSC1")  # Specify projects or NULL for all
)

# Connect 
con <- connectDROMADatabase("path/to/droma.sqlite")

# Check database structure
tables <- listDROMADatabaseTables()

Actually, there is no need for you to do these trivialities, just download the lastest version of DROMA_DB:10.5281/zenodo.15055905

Then use DROMA.Set package and DROMA_R to play with DROMA.

• Data Tables: {project}_{datatype} (e.g., CCLE_mRNA, gCSI_drug)
• Annotation Tables: sample_anno, drug_anno
• Metadata: projects table with project information

• mRNA: Gene expression data
• cnv: Copy number variation data
• mutation_gene: Gene-level mutation data
• mutation_site: Site-specific mutation data
• fusion: Gene fusion data
• meth: DNA methylation data
• proteinrppa: Reverse-phase protein array data
• proteinms: Mass spectrometry proteomics data

• drug: Drug sensitivity/response data
• drug_dose/viability: drug sensitivity raw data allows users to recalculate sensitivity metrics (IC50, AUC, AAC) and generate dose-viability plots.

• CellLine: Cancer cell lines
• PDO: Patient-derived organoids
• PDX: Patient-derived xenografts
• PDC: Patient-derived cells
• Clinical: Clinical patients data labeled as response or non-response, and needs special API to access from CTRDB database currently.

• sample_anno: Annotation data for samples, example:

• drug_anno: Annotation data for drugs, example:

The workflow/ directory contains comprehensive examples and tutorials:

1. 01-CreateDROMA.R - Basic database creation script

   # Simple database creation from all projects
   createDROMADatabase(db_path = "sql_db/droma.sqlite",
                       rda_dir = "data", 
                       projects = NULL)

2. 02-UpdateDROMA01.R - Database update script for adding new data

3. 02-UpdateDROMA02.Rmd - Comprehensive update workflow with documentation

4. 03-using_droma_database.R - Complete usage examples including:

   • Connecting to database and listing tables
   • Retrieving specific gene expression data (BRCA1 from CCLE/gCSI)
   • Filtering by data type (PDX models only)
   • Filtering by tumor type (breast cancer cell lines)
   • Proper connection management

5. README_SQL_DATABASE.md - Detailed SQL database functionality guide with:

   • Performance comparison vs traditional loading
   • Advanced usage patterns
   • Best practices and optimization tips

# Example 1: Get BRCA1 expression from specific sources
brca1_data <- getFeatureFromDatabase(
  select_feas_type = "mRNA",
  select_feas = "BRCA1",
  data_sources = c("CCLE", "gCSI")
)

# Example 2: Filter by model system
drug_pdx <- getFeatureFromDatabase(
  select_feas_type = "drug",
  select_feas = "Tamoxifen",
  data_type = "PDX"
)

# Example 3: Filter by tumor type
tp53_mutations <- getFeatureFromDatabase(
  select_feas_type = "mutation_gene",
  select_feas = "TP53",
  data_type = "CellLine",
  tumor_type = "breast cancer"
)

For detailed examples and best practices, refer to the workflow/ directory.

DROMA_DB/
├── data/                    # Input Rda files
│   ├── anno.Rda            # Sample and drug annotations
│   ├── mRNA.Rda            # Gene expression data
│   ├── drug.Rda            # Drug response data
│   ├── cnv.Rda             # Copy number data
│   └── ...
├── function/               # Database creation functions
│   └── function.R          # Main functions
├── workflow/               # ⭐ Example workflows and tutorials
│   ├── 01-CreateDROMA.R    # Database creation script
│   ├── 02-UpdateDROMA01.R  # Database update script
│   ├── 02-UpdateDROMA02.Rmd # Comprehensive update workflow
│   ├── 03-using_droma_database.R # Usage examples
│   └── README_SQL_DATABASE.md # SQL functionality guide
├── sql_db/                 # Output database directory
│   └── droma.sqlite        # Generated SQLite database
└── README.md              # This file

• Indexed Tables: All feature tables have indexes on feature_id for fast lookups
• Efficient Storage: Optimized SQLite schema reduces storage requirements
• Batch Processing: Bulk data insertion for improved performance
• Memory Management: Streaming data processing for large datasets

• R: Version 4.0.0 or higher
• Required Packages:
  • RSQLite (≥ 2.2.0)
  • DBI (≥ 1.1.0)
  • tools (base R)

Li, S., Peng, Y., Chen, M. et al. Facilitating integrative and personalized oncology omics analysis with UCSCXenaShiny. Commun Biol 7, 1200 (2024). https://doi.org/10.1038/s42003-024-06891-2

• DROMA_Set Package: https://github.com/mugpeng/DROMA_Set
• Data Repository: https://zenodo.org/records/15742800
• Documentation: Refer to workflow/ directory for detailed examples

This project is licensed under the Mozilla Public License 2.0 - see the LICENSE file for details.

For questions and support:

• Open an issue on GitHub
• Check the workflow examples in workflow/
• Review the function documentation in function/function.R

Documentation and Maintenance:

• Overhauled README.md: The project now features a comprehensive README with a detailed overview, core features, clear installation instructions, and practical usage examples.
• Streamlined Workflow: Removed deprecated preprocessing and database update scripts to simplify project maintenance.
• Linked Primary Datasource: Added a direct link to the Input data for DROMA_DB on Zenodo for improved transparency and accessibility.

Data and Feature Enhancements:

• Expanded Project Coverage: Integrated three new data sources: LICOB (PDOs), HKUPDO (PDOs), and CTRDB (clinical records), significantly broadening the database's scope.
• Enabled Raw Data Analysis: Introduced raw drug-dose response data for several projects. This new feature allows users to recalculate sensitivity metrics (IC50, AUC, AAC) and generate dose-viability plots.
• Standardized Sensitivity Scoring: Rescaled all drug sensitivity values to a unified 0-1 range where higher values indicate greater sensitivity. This involved converting UMPDO data with a rescaled 1 - IC50 and applying an AAC calculation based on tumor volume for Xeva data, as detailed in its official vignette.

Note: This database creation tool is designed to work seamlessly with the DROMA_Set package for comprehensive omics data analysis.

DROMA_DB - as the foundation for the broader DROMA ecosystem 🧬💊