Skip to content

CHANGELOG.md

Latest commit

 

History

History
281 lines (204 loc) · 8.36 KB

CHANGELOG.md

File metadata and controls

281 lines (204 loc) · 8.36 KB

RAPTOR Changelog

All notable changes to RAPTOR (RNA-seq Analysis Pipeline Testing and Optimization Resource) will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[2.1.1] - 2025-12-15

Feature Release - Adaptive Threshold Optimizer

This release introduces the Adaptive Threshold Optimizer (ATO), a data-driven approach to selecting significance thresholds for differential expression analysis. No more arbitrary cutoffs!

Highlights

# Verify ATO is available
python -c "from raptor.threshold_optimizer import optimize_thresholds; print('✅ ATO Ready!')"

Added

Adaptive Threshold Optimizer (ATO) - Major New Feature!

  • Data-Driven Threshold Selection - Replace arbitrary thresholds with scientifically justified values

    • Multiple p-value adjustment methods (BH, BY, Storey q-value, Holm, Hochberg, Bonferroni)
    • Five logFC optimization methods (MAD, mixture model, power-based, percentile, consensus)
    • π₀ estimation for true null proportion (Storey, Pounds & Cheng, histogram methods)
    • Analysis goal presets (discovery, balanced, validation)
    • Automatic threshold reasoning and explanation

  • Publication-Ready Output

    • Auto-generated methods text for papers
    • Comprehensive threshold comparison heatmaps
    • Volcano plots with optimized thresholds
    • P-value and logFC distribution visualizations
    • Export to CSV/Excel with full statistics

  • Dashboard Integration

    • New "🎯 Threshold Optimizer" page in interactive dashboard
    • Upload DE results (DESeq2/edgeR/limma compatible)
    • Demo data generation for testing
    • Interactive visualizations with Plotly
    • Download buttons for results and methods text

Dashboard Updates

  • Added Threshold Optimizer page (7th navigation page)
  • Added ATO availability indicator in sidebar
  • Added "What's New in v2.1.1" banner on home page
  • Updated navigation structure
  • Added session state management for ATO

Configuration Updates

  • New threshold_optimizer section in all config files
  • Updated config.yaml with full ATO documentation
  • Added use_adaptive_thresholds option to statistics section
  • Updated cloud container images to v2.1.1
  • Added ATO settings to publication and ensemble configs

New Module

raptor/threshold_optimizer/
├── __init__.py          # Module exports
├── ato.py               # AdaptiveThresholdOptimizer class
└── visualization.py     # Plotting functions

Main Classes:

  • AdaptiveThresholdOptimizer - Core optimization class
  • ThresholdResult - Named tuple for results
  • optimize_thresholds() - Convenience function

ATO Features

Feature Description
Analysis Goals discovery (permissive), balanced (standard), validation (stringent)
P-value Methods Benjamini-Hochberg, Benjamini-Yekutieli, Storey q-value, Holm, Hochberg, Bonferroni
LogFC Methods Auto (consensus), MAD-based, Mixture model, Power-based, Percentile
π₀ Estimation Storey's spline, Pounds & Cheng, Histogram-based
Visualizations Volcano, distributions, heatmaps, optimization summary

Changed

  • Updated version to 2.1.1 across all files
  • Enhanced __init__.py with ATO imports and availability flags
  • Updated launch_dashboard.py with ATO check on startup
  • All example configs updated for v2.1.1
  • Updated container image references to 2.1.1
  • Added use_adaptive_thresholds: true as recommended default

New Documentation

  • THRESHOLD_OPTIMIZER.md - Comprehensive ATO documentation

Fixed

  • Dashboard now gracefully handles missing ATO module
  • Improved error messages for threshold optimization failures
  • Fixed config validation for new threshold_optimizer section

⚙️ Dependencies

No new required dependencies. ATO uses existing scipy, numpy, and pandas.

Migration from v2.1.0

Full backward compatibility maintained!

# Just update the package
pip install --upgrade raptor-rnaseq

Enable ATO (Optional):

threshold_optimizer:
  enabled: true
  goal: "discovery"

Quick ATO Example

from raptor.threshold_optimizer import AdaptiveThresholdOptimizer
import pandas as pd

# Load DE results
df = pd.read_csv('deseq2_results.csv')

# Optimize thresholds
ato = AdaptiveThresholdOptimizer(df, logfc_col='log2FoldChange', pvalue_col='pvalue')
result = ato.optimize(goal='discovery')

print(f"Optimal logFC threshold: {result.logfc_threshold:.3f}")
print(f"Significant genes: {result.n_significant}")
print(f"\nMethods text:\n{result.methods_text}")

[2.1.0] - 2025-06-12

Major Release - ML Intelligence, Interactive Dashboard & PyPI Publication

This release represents a significant evolution of RAPTOR, introducing artificial intelligence, interactive visualization, cloud computing capabilities, and official PyPI publication while maintaining full backward compatibility with v2.0.0.

Published to PyPI

RAPTOR is now available on the Python Package Index!

# Install from PyPI
pip install raptor-rnaseq

# With all features
pip install raptor-rnaseq[all]

PyPI Page: https://pypi.org/project/raptor-rnaseq/

Added

Machine Learning System

  • ML-Based Pipeline Recommendations - Intelligent pipeline selection using machine learning
    • Random Forest model trained on 10,000+ real-world RNA-seq analyses
    • 85-90% accuracy in pipeline recommendations
    • Confidence scoring for all predictions
    • Model explainability with feature importance
    • Custom model training for lab-specific optimization

Interactive Dashboard

  • Web-Based Dashboard - Modern, interactive interface built with Streamlit
    • Zero-coding user interface for all RAPTOR features
    • Real-time analysis monitoring
    • Interactive quality control visualizations
    • Pipeline comparison plots
    • Export publication-ready figures

Advanced Quality Assessment

  • Comprehensive QC Module - Enhanced quality control and data assessment
    • Multi-level quality scoring (0-100 scale)
    • Automated contamination detection
    • Batch effect identification

Resource Monitoring

  • Real-Time Resource Tracking - Live monitoring of computational resources
    • CPU usage per pipeline
    • Memory consumption tracking
    • Cost estimation (cloud deployments)

Ensemble Analysis

  • Multi-Pipeline Ensemble - Consensus building across multiple pipelines
    • Weighted averaging of results
    • Confidence scoring per gene
    • Publication-quality ensemble reports

Parameter Optimization

  • Automated Parameter Tuning - Intelligent parameter optimization
    • Grid search and Bayesian optimization
    • Integration with ML recommendations

Automated Reporting

  • Publication-Ready Reports - Comprehensive automated documentation
    • HTML interactive reports
    • PDF static reports
    • Methods section generation

Cloud Integration

  • Multi-Cloud Support - Native cloud computing integration
    • AWS Batch, GCP, Azure support
    • Spot/preemptible instance support
    • Auto-scaling capabilities

🔧 Changed

  • Refactored configuration system for better flexibility
  • Enhanced error handling and recovery
  • Simplified installation process (now pip install raptor-rnaseq)
  • Updated Salmon, Kallisto, STAR support

Fixed

  • Fixed memory leak in long-running analyses
  • Corrected race condition in parallel processing
  • Fixed crash with non-standard chromosome names

Performance

  • 25% faster pipeline execution
  • 40% reduction in memory usage
  • Reduced startup time by 60%

[2.0.0] - 2024-05-15

Major Release - Initial Public Release

Added

  • Multi-pipeline RNA-seq analysis framework
  • Support for Salmon, Kallisto, STAR, RSEM, HTSeq
  • Automated quality control
  • Pipeline comparison metrics
  • Comprehensive configuration system

[1.0.0] - 2023-12-01

Initial Development Release

  • Basic framework structure
  • Support for 2 pipelines (Salmon, STAR)
  • Simple configuration

Version History

  • v2.1.1: Adaptive Threshold Optimizer (Current)
  • v2.1.0: ML Intelligence, Dashboard & PyPI
  • v2.0.0: Initial Public Release
  • v1.0.0: Development Release

How to Update

pip install --upgrade raptor-rnaseq

Author: Ayeh Bolouki
License: MIT
PyPI: https://pypi.org/project/raptor-rnaseq/