How listerlab/home works
Main Components
a) Input Processing Layer:
• File Input Handlers (supporting both custom format and BSSeeker2 CGmap files)
• Sample Configuration Parser
• Data Validation Components
b) Core Processing Components:
• Histogram Feature Generator
• SVM Classifier Module
• DMR Detection Engine (split into two main paths):
• Pairwise DMR Detection
• Time Series DMR Detection
c) Data Storage:
• Model Storage (saved_model directory)
• Training Data Repository
• Temporary Processing Storage
Component Relationships
[Input Files] → [Input Processing Layer]
• ↓
[Training Data] → [Feature Generation] → [SVM Classification]
• ↓
• [DMR Detection Engine]
• / \
• [Pairwise Mode] [Time Series Mode]
• ↓ ↓
• [DMR Output] [DMR Output]
Important Details
a) Processing Parameters:
• Configuration options for both modes
• Parameters (scorecutoff, pruncutoff, etc.)
b) Parallel Processing:
• Multi-core processing capability
• Chromosome-level parallelization option
Technical Specifications
A technical stack sidebar showing:
• Python 2.7
• Key Dependencies:
• numpy
• pandas
• scipy
• scikit-learn
• statsmodels