CountDataSet-class: Class "CountDataSet" - a container for count data from HTS...

In DESeq: Differential gene expression analysis based on the negative binomial distribution

Description

This is the main class for the present package.

Objects from the Class

Objects should be created with calls to newCountDataSet (q.v.).

Extends

Class eSet (package 'Biobase'), directly. Class VersionedBiobase (package 'Biobase'), by class "eSet", distance 2. Class Versioned (package 'Biobase'), by class "eSet", distance 3.

Note

Note: This is a summary for reference. For an explanation of the actual usage, see the vignette.

A CountDataSet object stores counts from an HTS data set and offers further slots which are populated during the analysis.

After creation with newCountDataSet, a CountDataSet typically contains a count table, i.e., a matrix of integer data, that is accessible with the accessor function counts. Each row of the matrix corresponds to a gene (or binding region, or the like), and each colum to an experimental sample. The experimental conditions of the samples are stored in a factor (with one element for each row of the counts matrix), which can be read with the accessor function conditions.

In the following analysis steps, further data slots are populated. First, the size factors can be estimated with estimateSizeFactors, which are afterwards accessible via sizeFactors. Then, the dispersions (variance fits) are estimated with estimateDispersions. The resulting estimates are stored in phenoData columns, accessible via pData, with the column names staring with disp_. The intermediate steps of the fit are stored in the environment-values slot fitInfo (see estimateDispersions for details).

Internally, the mentioned data is stored in slots as follows:

As CountDataSet is derived from eSet, it has a phenoData slot which allows to store sample annotation. This is used to store the factor with the conditions, as a data frame column named condition, and to store the size factors, as an numeric data frame column named sizeFactor. If the user creates an object with multivariate design, i.e., passes a data frame instead of a factor for conditions, this data frame's columns are placed in the phenoData slot instead of the condition column. Furthermore, the function estimateDispersions adds columns with the dispersion values to be used by nbinomTest and fitNbinomGLMs. These columns have names starting with disp_.

The user may add further columns to the phenoData AnnotatedDataFrame.

The counts table is stored in the eSet's assayData locked environment with the name counts.

The slot dispInfo is an environment containing lists, one for each set of estimated dispersion values and the slot dispTable (with accessor dispTable shows the assignment of conditions to dispersion estimates. See estimateDispersions

Examples

# See the vignette

DESeq documentation built on April 28, 2020, 6:37 p.m.