addDesignMatrix: Add Design Matrix

In BussemakerLab/SelexGLM: Functions for performing feature-based analyis of SELEX-seq data using generalized linear models

Description

Uses top binding windows to align probes and add corresponding design matrix to topModelMatch for regression based on feature design in model object.

Usage

addDesignMatrix(data, model)

Arguments

data	Table output by topModelMatch containing probes, probe counts and information about the top binding view for each probe.
model	Object of class model used to build design matrix from the output of topModelMatch.

Details

Builds a design matrix for probes in data using the features contained in model. The following features will be added for use in the glm regression:

1. logProb: log(data$Probability)

2. N.Ax, N.Cx, N.Gx, N.Tx: A nucleotide at position x is represented by a set of 4 nucleotide variables that are equal to 0 or data$Round depending on the identity of the nucleotide in question. If reverse complement symmetry is being used, contributions from x and its reverse complement symmetric equivalent nucleotide, x', are both represented by a single block of 4 nucleotides.

3. fixedNddG: If useFixedValuesOffset.N == TRUE, fixedNddG is the ddG contribution from nucleotides not included in model@features@[e-mail].

4. Round.R: The round of a probe, r, is represented by a value of 1 in column Round.r and 0 in all other round columns.

5. Strand.[FR]: If includeDNAstrand == TRUE in model input, the DNA strand of the top view is represented by (data$Strand.F, data$Strand.R) equal to 0 or data$Round depending on which strand the top view is located on.

6. Strand.[FR]v: If includeView == TRUE in model input, the View and strand of the top view is represented by a set of variables data$Strand.[FR]v, where v is in [1, (number of views per strand)]. These variables are equal to data$Round if the top view is the one represented by that variable, or zero if not.

7. Shape.[shape]x: A shape parameter at position x is represented by a single variable Shape.[shape]x, which is equal to the shapeTable lookup value for the shape parameter at position x. A variable of this kind will be included for each shape in model@shapeParamsUsed[[1]] and each position in model@features@[e-mail]

Value

Data frame containing probes with design matrix built for regression input.

BussemakerLab/SelexGLM documentation built on May 17, 2019, 5:41 p.m.