DEVis (version 1.0.1)

de_density_plot: Visualize density plots of fold-change or significance values for aggregated data sets.

Description

This function plots log2 fold-change or adjusted p-values for all differentially expressed genes for each contrast in a result set. Data aggregation across a series of result sets entails that not every gene in an aggregated data set will necessarily be differentially expressed, depending on the aggregation method. For instance, a gene that is differentially expressed in a day1 contrast that is not significant at day2 will be included in a union based aggregation of genes for day1 and day2. Visualizing the density of fold-change or p-values can reveal to what extent aggregation was consistent across conditions and can inform decision making in data aggregation.

Usage

de_density_plot(res_list, filename = "de_density_plot.pdf",
  type = "pval", method = "union", returnData = FALSE)

Arguments

res_list

A list of DESeq result sets. Results can be calculated individually using DESeq's results() function. Lists of results can be created by creating a list(result1, result2 ... result_N).

filename

Filename for output plot. Valid extensions are ".pdf" and ".png". File generation can be turned off using set_output_mode("screen"). Output will be written to the /DE/density_plots/ directory.

type

The type of data to be displayed in this plot. Valid selections are "lfc" (log2foldChange) and "pval".

method

The method for computing overlaying results. Valid selections are "union" or "intersection". Union merges data for all result sets for genes that are differentially expressed in at least 1 result set. Intersection merges data for genes that only are differentially expressed in all result sets provided.

returnData

Boolean. Determines if this visualization should return data used to generate the visualization. Default=FALSE.

Value

If returnData is true, this function will return a data frame for aggregated differentially expressed genes containing gene names and log2 fold-change or adjusted p-values relative to the experimental control condition.

Examples

# NOT RUN {
 #This example assumes an experimenal design of ~Condition_Time.

#Prepare a result list.
res.day1 <- results(dds, contrast=c("Condition_Time", "day1_disease", "day1_control"))
res.day2 <- results(dds, contrast=c("Condition_Time", "day2_disease", "day2_control"))
res.day3 <- results(dds, contrast=c("Condition_Time", "day3_disease", "day3_control"))
myResList <- list(res.day1, res.day2, res.day3)

/*
 * Aggregate data for all contrasts in the result list using union aggregation.
 * Display the density plot of p-values for the aggregated data.
 */
de_density_plot(res_list=myResList, filename="DE_density_union_pval.pdf",
                 type="pval", method="union", returnData=FALSE)

/*
 * Aggregate data for all contrasts in the result list using intersection aggregation.
 * Display the density plot of log fold-change values for the aggregated data.
 * Store the aggregate data as DE_lfc_intersect_df.
 */
DE_lfc_intersect_df <- de_density_plot(res_list=myResList,
                                        filename="DE_density_union_pval.pdf",
                                        type="lfc", method="intersection",
                                        returnData=TRUE)

# }