de_filter: Apply a custom fold-change filter to an aggregated data...

In DEVis: A Differential Expression Analysis Toolkit for Visual Analytics and Data Aggregation

Description

This function allows you to filter an aggregated master result to only contain genes with fold-changes less than or greater than a specified threshold for a given metric. Several metrics are provided for flexibility of selection. The use of metrics in this function makes it possible to ask questions such as, "which genes have a mean fold-change of at least 2?", "which genes have a fold change of less than 5?", or "which genes have fold-change with a variance across all conditions of at least 10?"

Usage

de_filter(master_df, metric, threshold, operator, absolute = TRUE)

Arguments

master_df	An aggregated master DE dataframe as produced using create_master_res().
metric	Metric to be used for filtering. All metrics are gene-wise calculated. That is, for each gene in the aggregate data set, a metric is calculated that is then used to filter the data. Valid metrics are: "max", "min", "mean", "variance", "sd". The "max" metric is the gene-wise maximum fold-change value. The "min" metric is the gene-wise minimum fold-change value. The "mean" metric is the gene-wise fold-change mean. The "variance" metric is the gene-wise variance. The "sd" metric is the gene-wise standard deviation.
threshold	The cutoff value to use for filtering. Numeric.
operator	Operator to apply to the filtering process. Options are: "less", "greater".
absolute	Use absolute values when applying filters to filter from both negative and positive ends of the fold change distribution. Boolean. Default=TRUE.

Value

This function returns a data frame that has been filtered by the specified criteria.

See Also

create_master_res

Examples

## Not run: 

#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)

#Prepare an aggregate master data frame.
aggregate_df <- create_master_res(res_list=myResList, filename="master_DE.txt")

/*
 * Filter the aggregate data to contain only genes whose gene-wise maximum
 * fold-change is greater than 10. This will return genes whose gene-wise maximum
 * is > 10.
 */
de_filtered <- de_filter(master_df=aggregate_df, metric="max", threshold=10,
                          operator="greater", absolute=FALSE)


/*
 * Filter the aggregate data to contain only genes whose gene-wise minimum
 * fold-change is greater than an absoulte value of 4.  This will return genes
 * where the gene-wise minimum is < -4 and > +4.
 */
de_filtered <- de_filter(master_df=aggregate_df, metric="min", threshold=4,
                          operator="greater", absolute=TRUE)


/*
 * Filter the aggregate data to contain only genes whose gene-wise standard
 * deviation is less than an absoulte value of 2.  This will return genes
 * where the gene-wise standard deviation is > -2 and < +2.
 */
de_filtered <- de_filter(master_df=aggregate_df, metric="sd", threshold=2,
                          operator="less", absolute=TRUE)


## End(Not run)

DEVis documentation built on May 2, 2019, 3:18 p.m.