Panel Merger Engine

The merger engine combines data from multiple sources and applies the scoring algorithm to determine final gene inclusion.

Overview

The PanelMerger class implements the core scoring logic:

• Data standardization - Ensures consistent data formats across sources
• Evidence aggregation - Combines evidence from multiple sources per gene
• Scoring algorithm - Applies configurable weights and thresholds
• Decision logic - Determines final gene inclusion with veto capabilities
• Quality metrics - Provides detailed scoring statistics

Scoring System

The merger implements a sophisticated scoring system that balances:

1. Source reliability - Evidence scores based on clinical trust
2. Classification quality - Multipliers for evidence classifications
3. Source diversity - Confidence scores based on supporting sources
4. Clinical priorities - Final weights and veto capabilities

API Reference

custom_panel.engine.merger

Gene panel merger and scoring engine.

This module provides functionality to merge gene panels from multiple sources, apply scoring logic, and make inclusion decisions.

PanelMerger

PanelMerger(config)

Gene panel merger and scoring engine.

Initialize the panel merger.

Parameters:

Name	Type	Description	Default
config	dict[str, Any]	Configuration dictionary	required

create_master_list

create_master_list(dataframes, output_manager=None)

Create master gene list from multiple source DataFrames.

Parameters:

Name	Type	Description	Default
dataframes	list[DataFrame]	List of standardized DataFrames from different sources	required
output_manager	Any	Optional output manager for saving intermediate files	None

Returns:

Type	Description
DataFrame	Master DataFrame with merged and scored genes

export_gene_lists

export_gene_lists(df, output_dir)

Export gene lists for different categories.

Parameters:

Name	Type	Description	Default
df	DataFrame	Master DataFrame	required
output_dir	str | Path	Output directory	required

Returns:

Type	Description
dict[str, str]	Dictionary mapping category to output file path

filter_by_category

filter_by_category(df, category)

Filter genes by inclusion.

Parameters:

Name	Type	Description	Default
df	DataFrame	Master DataFrame	required
category	str	Category to filter by (only "include" is supported now)	required

Returns:

Type	Description
DataFrame	Filtered DataFrame

get_scoring_summary

get_scoring_summary(df)

Generate summary statistics for scoring results.

Parameters:

Name	Type	Description	Default
df	DataFrame	Scored DataFrame	required

Returns:

Type	Description
dict[str, Any]	Summary statistics dictionary