EDEG (JDEG) Analysis

DOLPHIN performs exon-level (junction-level) differential gene expression analysis using MAST results generated by the Seurat toolkit.

The MAST method [1] is used to evaluate exon-level statistical significance, and exon weights are integrated to compute gene-level scores.

[1]

Finak et al. (2015). MAST: a flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data. Genome Biology. https://doi.org/10.1186/s13059-015-0844-5

[2]

Satija et al. (2015). Seurat: R toolkit for single cell genomics. https://satijalab.org/seurat/

DOLPHIN.EDEG.call_convert.run_h5ad_rds(input_anndata, output_rds)

Convert a .h5ad file to a Seurat RDS object by calling an R script.

Parameters:

• input_anndata (str) – Path to the input .h5ad file to be converted.

• output_rds (str) – Path where the resulting .rds file will be saved.

Returns:

The Seurat object is saved to the location specified by output_rds.

Return type:

None

DOLPHIN.EDEG.generate_EDEG.run_edeg(seurat_output, adata_input, gtf_path, output)

Aggregate exon-level marker results into gene-level statistics using weighted methods.

This function converts exon-level marker results from Seurat using MAST into gene-level statistical insights. It applies the Stouffer method to combine exon-level p-values, weighted by exon length, and computes average absolute log2 fold changes, also weighted by exon length.

Parameters:

• seurat_output (str) – Path to the CSV file containing exon-level differential expression results from Seurat.

• adata_input (str) – Path to the .h5ad file used for generating exon-level data.

• gtf_path (str) – Path to the GTF-derived annotation file generated by DOLPHIN.

• output (str) – Path where the final gene-level results will be saved as a CSV file.

Returns:

A DataFrame containing gene-level statistics including: - MAST_weighted_abs_avg_log2FC: Weighted average of absolute log2 fold changes per gene. - MAST_weighted_stouffer_pval: Combined p-value using the weighted Stouffer method. - MAST_weighted_stouffer_pval_adj_bonf: Bonferroni-adjusted Stouffer p-value.

Return type:

pd.DataFrame

DOLPHIN.EDEG.generate_JDEG.run_jdeg(seurat_output, output)

Aggregate junction-level marker results into gene-level statistics using stouffer method.

This function converts junction-level marker results from Seurat using MAST into gene-level statistical insights. It applies the Stouffer method to combine junction-level p-values, and computes average absolute log2 fold changes.

Parameters:

• seurat_output (str) – Path to the CSV file containing exon-level differential expression results from Seurat.

• output (str) – Path where the final gene-level results will be saved as a CSV file.

Returns:

A DataFrame containing gene-level statistics including: - MAST_abs_avg_log2FC: Average of absolute log2 fold changes per gene. - MAST_stouffer_pval: Combined p-value using the Stouffer method. - MAST_stouffer_pval_adj_bonf: Bonferroni-adjusted Stouffer p-value.

Return type:

pd.DataFrame