coupling_test.py

​

High-level description

This file contains unit tests for the coupling estimators implemented in cassiopeia/tools/coupling.py, specifically focusing on the compute_evolutionary_coupling function. These tests ensure the correctness of the evolutionary coupling calculations under various scenarios, including different input parameters and data configurations.

​

Code Structure

The code consists of a single test class, TestDataUtilities, which houses multiple test methods. Each test method focuses on a specific aspect of the compute_evolutionary_coupling function, such as its basic functionality, handling of custom dissimilarity maps, minimum proportion thresholds, and potential error cases.

​

Symbols

​

TestDataUtilities

​

Description

This class contains unit tests for data utility functions, specifically focusing on the compute_evolutionary_coupling function.

​

Inputs

• None

​

Outputs

• None

​

Internal Logic

This class utilizes the unittest framework to define and execute test methods. Each test method sets up a specific scenario using a predefined tree structure and cell metadata, then calls the compute_evolutionary_coupling function with specific parameters. The expected output is then compared to the actual output to verify the correctness of the function.

test_evolutionary_coupling_basic

• Description: Tests the basic functionality of the compute_evolutionary_coupling function with default parameters.
• Inputs: None
• Outputs: None
• Internal Logic:
  1. Sets up a CassiopeiaTree with predefined tree topology and cell metadata.
  2. Calls compute_evolutionary_coupling with the tree, “CellType” as the meta variable, a fixed random state, minimum proportion of 0, and 10 shuffles.
  3. Computes the expected inter-cluster distances and evolutionary coupling matrix based on the predefined data and parameters.
  4. Asserts that the calculated evolutionary coupling matrix matches the expected matrix.
  5. Asserts that a CassiopeiaError is raised when attempting to compute evolutionary coupling with a numerical meta variable (“nUMI”).

test_evolutionary_coupling_custom_dissimilarity_map

• Description: Tests the functionality of compute_evolutionary_coupling with a custom dissimilarity map.
• Inputs: None
• Outputs: None
• Internal Logic:
  1. Sets up a CassiopeiaTree and a custom weight matrix representing the dissimilarity map.
  2. Calls compute_evolutionary_coupling with the tree, “CellType” as the meta variable, a fixed random state, minimum proportion of 0, 10 shuffles, and the custom dissimilarity map.
  3. Computes the expected inter-cluster distances and evolutionary coupling matrix based on the predefined data, parameters, and custom dissimilarity map.
  4. Asserts that the calculated evolutionary coupling matrix matches the expected matrix.

test_evolutionary_coupling_minimum_proportion

• Description: Tests the functionality of compute_evolutionary_coupling with a minimum proportion threshold for categories.
• Inputs: None
• Outputs: None
• Internal Logic:
  1. Sets up a CassiopeiaTree and modifies the cell metadata to include a category (“TypeD”) that falls below the minimum proportion threshold.
  2. Calls compute_evolutionary_coupling with the tree, “CellType” as the meta variable, a fixed random state, a minimum proportion of 1/6 (which excludes “TypeC” and “TypeD”), and 10 shuffles.
  3. Asserts that the resulting evolutionary coupling matrix does not include the excluded categories (“TypeC” and “TypeD”).
  4. Computes the expected inter-cluster distances and evolutionary coupling matrix based on the remaining categories and asserts that the calculated matrix matches the expected matrix.