graph LR
deseq2_norm["deseq2_norm"]
deseq2_norm_fit["deseq2_norm_fit"]
deseq2_norm_transform["deseq2_norm_transform"]
deseq2_norm -- "calls" --> deseq2_norm_fit
deseq2_norm -- "calls" --> deseq2_norm_transform
The Data Normalization subsystem in pydeseq2 is primarily composed of three interconnected functions, rather than a complex class hierarchy. These functions collectively implement the DESeq2 median-of-ratios normalization method.
This function serves as the primary public interface for performing DESeq2 median-of-ratios normalization. It orchestrates the entire normalization workflow by internally calling deseq2_norm_fit to compute sample-specific size factors and then deseq2_norm_transform to apply these factors, returning the normalized count data. It encapsulates the complete normalization process for user convenience.
Related Classes/Methods:
deseq2_norm(1:1)
This function is responsible for the "fitting" phase of the DESeq2 normalization. It takes raw count data as input and calculates the sample-specific size factors. These factors are crucial for adjusting for differences in sequencing depth and other technical variations between samples, forming the basis for comparable downstream analyses.
Related Classes/Methods:
deseq2_norm_fit(1:1)
This function performs the "transformation" phase of the DESeq2 normalization. It takes the raw count data and the previously computed size factors (obtained from deseq2_norm_fit). It then divides the raw counts by their respective size factors to produce the final normalized count matrix, making the data comparable across samples.
Related Classes/Methods:
deseq2_norm_transform(1:1)