sleuth_kruskal_wallis: Run a Kruskal-Wallis One-Way ANOVA on non-parametric data at...

In nodrogluap/gumshoe: Supplemental Statistical Tests and Utilities for Kallisto/Sleuth RNASeq Analysis

Run a Kruskal-Wallis One-Way ANOVA on non-parametric data at the isoform level for a selected gene. The quantile of data the test is run upon is calculated from the the est_counts of all the data and the resulting value is used to filter based upon the mean of the est_counts for each individual transcript. For example, for a quantile probability of 0.25, the function will determine the quantile probability for 0.25 and assume this value to be 50. The function will then determine the mean est_counts for each transcript of the gene and if the mean is greater than the quantile calculated value of 50, it will be included when running the test.

Description

Run a Kruskal-Wallis One-Way ANOVA on non-parametric data at the isoform level for a selected gene. The quantile of data the test is run upon is calculated from the the est_counts of all the data and the resulting value is used to filter based upon the mean of the est_counts for each individual transcript. For example, for a quantile probability of 0.25, the function will determine the quantile probability for 0.25 and assume this value to be 50. The function will then determine the mean est_counts for each transcript of the gene and if the mean is greater than the quantile calculated value of 50, it will be included when running the test.

Usage

sleuth_kruskal_wallis(sleuth_obj, gene, grouping, iqf = 0, threshold = 0.05)

Arguments

sleuth_obj	An existing Sleuth object as generated by sleuth_prep() and fit by sleuth_fit()
gene	The name of a gene as a string to asses the isoform expression level between samples
grouping	A column name in the sample to covariate data frame provided when running sleuth_prep / sleuth_interpret
iqf	A numeric value between 0 and 1 (inclusive) to denote the quantile of data to run the test upon.
threshold	A numeric value for the p-value cutoff when running the resulting pairwise comparison.

Value

Results of a kruskal-wallis test at the isoform level

Examples

# Given a Sleuth object, run the Kruskal-Wallis test of the est_counts factor between isoform groups for all samples.
sleuth_kruskal_wallis(so_nominal, "ABH1", "sex", iqf = 0.25)

nodrogluap/gumshoe documentation built on Feb. 28, 2023, 6:15 p.m.