de-quack 0.1.1

Python library for differential expression data storage and visualization using DuckDB.

de_quack

Python/R library for differential expression (DE) data management and visualization using DuckDB.

Stores DE analysis results with automatic gene annotation, enabling reproducible and efficient querying of experimental data.

Features

• Efficient Storage: DuckDB-backed persistence for gene expression results
• Automatic Annotation: Maps gene symbols and Ensembl IDs with canonical references
• Flexible Querying: Filter results using simple keyword arguments (e.g., padj__lt=0.05)
• Multi-language Support: Native Python and R interfaces
• Visualization: Built-in volcano plot generation with publication-ready styling
• Data Normalization: Handles variable column names and formats automatically

Installation

pip install de_quack

For Excel export support (optional):

pip install de_quack[excel]

Quick Start

Python

from de_quack import de_quackling, volcano_plot

with de_quackling('results.duckdb') as db:
    db.initialize_gene_table('human')
    db.insert_to_database('data.txt')
    results = db.query('gene_results', padj__lt=0.05)
    print(results)
    
# Generate volcano plot
volcano_plot(results, padj=0.05, log2fc=1, plot_file='volcano.png')

R

# Find and source the wrapper directly from the pip installation directory
py_pkg_path <- dirname(reticulate::import("de_quack")$`__file__`)
source(file.path(py_pkg_path, "wrapper.R"))

# Initialize
duck <- de_quack(db_path = "results.duckdb")
duck$connect()
duck$initialize_gene_table('human')
duck$insert_to_database('data.txt')
results <- duck$query('gene_results', padj__lt=0.05)
duck$volcano_plot(results, padj=0.05, log2fc=1, plot_file='volcano.png')
duck$close()

Database Schema

Tables

• experimental_data: Metadata about DE analysis experiments
• gene_results: Individual gene-level results with statistics
• genes: Reference gene annotations (human, etc.)

Advanced Querying & JSON Fallback

DeDuck supports intuitive field modifiers (__lt, __gt, __lte, __gte, __ne). If a metric isn't a native column in the database, DeDuck automatically queries it out of custom nested JSON metadata dynamically:

# Regular column filtering + automatic nested JSON key metadata extraction
with de_quackling as db:
    results = db.query('gene_results', padj__lt=0.05, custom_biotype__ne='pseudogene')

License

MIT