Skip to content

Latest commit

 

History

History
50 lines (29 loc) · 3.35 KB

File metadata and controls

50 lines (29 loc) · 3.35 KB
graph LR
    DeseqStats["DeseqStats"]
    DeseqDataSet["DeseqDataSet"]
    Inference["Inference"]
    DeseqStats -- "uses" --> DeseqDataSet
    DeseqDataSet -- "provides data to" --> DeseqStats
    DeseqStats -- "delegates to" --> Inference
    Inference -- "performs calculations for" --> DeseqStats
    click Inference href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/PyDESeq2/Inference.md" "Details"
Loading

CodeBoardingDemoContact

Details

The Differential Expression Analysis component is primarily orchestrated by the DeseqStats class, which leverages data from DeseqDataSet and statistical capabilities provided by the Inference component (specifically DefaultInference). These three components are fundamental to the subsystem's operation.

DeseqStats

This is the central orchestrator of the differential expression analysis workflow. It initializes with a DeseqDataSet object and a contrast, then provides methods to run Wald tests, perform LFC shrinkage, apply outlier filtering, and adjust p-values. Its summary() method consolidates these steps to produce the final differential expression results. It delegates core statistical computations to an Inference implementation.

Related Classes/Methods:

DeseqDataSet

This component serves as the primary data container for the differential expression analysis. It encapsulates raw count data, sample metadata, the design matrix, and stores intermediate results such as size factors, dispersions, and estimated LFCs. DeseqStats relies on a pre-processed DeseqDataSet object to perform its analysis.

Related Classes/Methods:

Inference [Expand]

The Inference component defines the interface for core statistical inference routines, while DefaultInference provides the concrete implementation. This component is responsible for performing the actual Wald tests and LFC shrinkage calculations. DeseqStats delegates these computationally intensive statistical tasks to an instance of this component, allowing for modularity and potential alternative statistical backends.

Related Classes/Methods: