calc_sc_gsea_score: Calculates scores for pre-ranked GSEA

Creates a vector of scores that can be used to rank genes in order to be submitted to GSEA

Usage

calc_sc_gsea_score(de_table)

Arguments

de_table

A differential expression table generated from the Seurat FindMarkers function. Contains the column headers p_val, avg_log2FC, pct.1, pct.2, and p_val_adj

Value

Returns a named vector of GSEA scores for genes

Details

The scoring metric uses a Seurat FindMarkers output table and incorporates the p-value and the fold change to determine magnitude, while using the percent expression values as weighting factors. The basic formula is:

-log_10(p-value) * sign (log2FC) * max(pct.1, pct.2) + log2FC

If the p-value is recorded as zero (i.e. below machine error), the formula is adjusted

500 * sign(log2FC) * max(pct.1, pct.2) + log2FC

This formula should preserve the ranking as indicated by magnitude of fold change,while also diminishing the significance of lowly-expressed genes.