Package {diffdriver}

Add intercept, if have functtypecode, then code and move to the front. different from driverMAPS! only allows functtypecode =7 ||8 when functypecode is included in selectvars.

Description

Add intercept, if have functtypecode, then code and move to the front. different from driverMAPS! only allows functtypecode =7 ||8 when functypecode is included in selectvars.

Usage

ddmcode(matrixlist, selectvars, functypecodelevel = NULL)

Arguments

diffDriver model

Description

This model is applied on data of a single gene. It will infer effect size for both sample-level variable and positional level functional annotations. We used an EM algorithm to infer parameters.

Usage

ddmodel(mut, e, mr, fe, label, ...)

Arguments

Value

A list with the following elements:

pvalue

The likelihood-ratio test p-value comparing the null model without a context effect to the alternative model with a context effect.

res.null

A list returned by the null-model fit. It contains loglikelihood, beta0, and alpha.

res.alt

A list returned by the alternative-model fit. It contains loglikelihood, beta0, and alpha.

Examples

# Synthetic mutation matrix: rows are positions and columns are samples.
mut <- matrix(0, nrow = 6, ncol = 20)

mut[1, 1] <- 1
mut[2, 5] <- 1

mut[c(1, 2), 11] <- 1
mut[c(2, 3), 12] <- 1
mut[c(3, 4), 13] <- 1
mut[c(4, 5), 14] <- 1
mut[c(5, 6), 15] <- 1
mut[c(1, 6), 16] <- 1
mut[c(1, 3), 17] <- 1
mut[c(2, 4), 18] <- 1
mut[c(3, 5), 19] <- 1
mut[c(4, 6), 20] <- 1

e <- c(rep(0, 10), rep(1, 10))

# Mutation rates and functional effects are on the log scale.
mr <- matrix(log(0.002), nrow = nrow(mut), ncol = ncol(mut))
fe <- rep(log(3), nrow(mut))

# Each row represents a separate mutation position.
label <- factor(seq_len(nrow(mut)))

result <- ddmodel(mut, e, mr, fe, label = label)
result$pvalue

diffDriver model only for binary phenotype.

Description

This function uses the model as cmodel.frac, but generalizes to take more than 1 functional categories. This model is applied on data of a single gene

Usage

ddmodel_binary(mut, e, bmr, fe)

Arguments

Value

A list with the following elements:

pvalue

The likelihood-ratio test p-value comparing the null and alternative binary-phenotype models.

null.eta0

The fitted null-model parameter.

alt.eta

The fitted alternative-model parameters for the two phenotype groups.

null.ll

The maximized log-likelihood under the null model.

alt.ll

The maximized log-likelihood under the alternative model.

all

A named numeric vector containing pvalue, null_para, e=0, e=1, null_likelihood, and alt_likelihood.

Examples

# Synthetic mutation matrix: rows are positions and columns are samples.
mut <- matrix(
  c(
    1, 0, 0, 0, 0, 0, 1, 0,
    0, 0, 0, 1, 0, 0, 0, 0,
    0, 1, 0, 0, 0, 0, 0, 1
  ),
  nrow = 3,
  byrow = TRUE
)
e <- c(0, 0, 0, 0, 1, 1, 1, 1)

# Background mutation rates and functional effects are on the log scale.
bmr <- matrix(log(0.001), nrow = nrow(mut), ncol = ncol(mut))
fe <- log(c(2, 2, 2))

result <- ddmodel_binary(mut, e, bmr, fe)
result$pvalue

diffDriver model only for binary phenotype, assuming bmr the same across samples.

Description

This function uses the model as cmodel.frac, but generalizes to take more than 1 functional categories. This model is applied on data of a single gene. This should give the same results as ddmodel_binary defined above.

Usage

ddmodel_binary_simple(mut, e, bmr, fe)

Arguments

Value

A list with the following elements:

pvalue

The likelihood-ratio test p-value comparing the null and alternative binary-phenotype models.

null.eta0

The fitted null-model parameter.

alt.eta

The fitted alternative-model parameters for the two phenotype groups.

null.ll

The maximized log-likelihood under the null model.

alt.ll

The maximized log-likelihood under the alternative model.

Examples

# Synthetic mutation matrix: rows are positions and columns are samples.
mut <- matrix(
  c(
    1, 0, 0, 0, 0, 0, 1, 0,
    0, 0, 0, 1, 0, 0, 0, 0,
    0, 1, 0, 0, 0, 0, 0, 1
  ),
  nrow = 3,
  byrow = TRUE
)
e <- c(0, 0, 0, 0, 1, 1, 1, 1)

# Background mutation rates and functional effects are on the log scale.
bmr <- matrix(log(0.001), nrow = nrow(mut), ncol = ncol(mut))
fe <- log(c(2, 2, 2))

result <- ddmodel_binary_simple(mut, e, bmr, fe)
result$pvalue

Run diffDriver with Input Files

Description

This function runs diffDriver.

Usage

diffdriver(
  gene,
  mut,
  pheno,
  anno_dir = ".",
  k = 6,
  totalnttype = 96,
  BMRmode = c("signature", "regular"),
  output_dir = tempdir(),
  output_prefix = "diffdriver_results"
)

Arguments

Chromosome Position Ref Alt SampleID
1 19 55653236 C T TCGA-N6-A4VE-01A-11D-A28R-08

SampleID SmokingCessation BMI
TCGA-N5-A4R8-01A-11D-A28R-08 0.5319630 20.0
TCGA-N5-A4RD-01A-11D-A28R-08 0.0448991 24.4

Value

A data frame with one row per tested gene. The returned columns include dd.p, mlr.p, mlr.v2.p, fisher.p, binom.p, and lr.p, which are p-values from DiffDriver and comparator methods; the corresponding .fdr columns, which contain Benjamini-Hochberg adjusted p-values; and mut.E1, mut.E0, E1, and E0, which summarize mutation and sample counts in the two phenotype groups used by the Fisher test.

Examples

phenof <- system.file(
  "extdata", "example_phenotypes.txt",
  package = "diffdriver"
)
genef <- system.file(
  "extdata", "example_gene.txt",
  package = "diffdriver"
)
mutf <- system.file(
  "extdata", "example_mutations.txt",
  package = "diffdriver"
)

pheno <- read.table(phenof, header = TRUE)
gene <- read.table(genef, header = FALSE)
mut <- read.table(mutf, header = TRUE)

## Not run: 
# Download the annotation files and replace this with their directory.
annodir <- "/path/to/annodir96"

result <- diffdriver(
  gene = gene,
  mut = mut,
  pheno = pheno,
  anno_dir = annodir,
  k = 6,
  totalnttype = 96,
  BMRmode = "signature",
  output_dir = tempdir(),
  output_prefix = "diffdriver_example"
)
result

## End(Not run)

gene level binomial test

Description

gene level binomial test

Usage

genebinom(mut, e)

Arguments

Value

A list with the following element:

pvalue

The p-value from a gene-level binomial test comparing mutation counts between the two phenotype groups.

Examples

# Synthetic mutation matrix: rows are positions and columns are samples.
mut <- matrix(
  c(
    1, 0, 0, 1, 0, 0,
    0, 1, 0, 0, 1, 0
  ),
  nrow = 2,
  byrow = TRUE
)
e <- c(0, 0, 0, 1, 1, 1)

result <- genebinom(mut, e)
result$pvalue

gene level fisher's exact test

Description

gene level fisher's exact test

Usage

genefisher(mut, e)

Arguments

Value

A list with the following elements:

pvalue

The p-value from a gene-level Fisher's exact test.

count

A numeric vector c(gmutc, gmutn, mu.c, mu.n), where gmutc and gmutn are mutation counts in the two phenotype groups, and mu.c and mu.n are the corresponding numbers of samples.

Examples

# Synthetic mutation matrix: rows are positions and columns are samples.
mut <- matrix(
  c(
    1, 0, 0, 1, 0, 0,
    0, 1, 0, 0, 1, 0
  ),
  nrow = 2,
  byrow = TRUE
)
e <- c(0, 0, 0, 1, 1, 1)

result <- genefisher(mut, e)
result$pvalue
result$count

gene level logistic regression

Description

gene level logistic regression

Usage

genelr(mut, e, covariates = rep(1, length(e)))

Arguments

Value

A list with the following elements:

res

The summary object from the fitted logistic regression model glm(e ~ dstatus + covariates, family = binomial(link = "logit")).

pvalue

The p-value for the mutation-status coefficient dstatus.

Examples

# Synthetic mutation matrix: rows are positions and columns are samples.
mut <- matrix(
  c(
    1, 0, 0, 1, 0, 0,
    0, 1, 0, 0, 1, 0
  ),
  nrow = 2,
  byrow = TRUE
)
e <- c(0, 0, 0, 1, 1, 1)

result <- genelr(mut, e)
result$pvalue

gene level multiple linear regression

Description

gene level multiple linear regression

Usage

mlr(mut, e, covariates = rep(1, length(e)))

Arguments

Value

A list with the following elements:

res

The summary object from the fitted linear model lm(e ~ dstatus + covariates).

pvalue

The p-value for the mutation-status coefficient dstatus.

Examples

# Synthetic mutation matrix: rows are positions and columns are samples.
mut <- matrix(
  c(
    1, 0, 0, 1, 0, 0,
    0, 1, 0, 0, 1, 0
  ),
  nrow = 2,
  byrow = TRUE
)
e <- c(0, 0, 0, 1, 1, 1)

result <- mlr(mut, e)
result$pvalue

gene level multiple linear regression, correcting for total number of mutations

Description

gene level multiple linear regression, correcting for total number of mutations

Usage

mlr.v2(mut, e, nmut, covariates = 1)

Arguments

Value

A list with the following elements:

res

The summary object from the fitted linear model lm(e ~ dstatus + nmut + covariates).

pvalue

The p-value for the mutation-status coefficient dstatus, after adjusting for total mutation count and covariates.

Examples

# Synthetic mutation matrix: rows are positions and columns are samples.
mut <- matrix(
  c(
    1, 0, 0, 1, 0, 0,
    0, 1, 0, 0, 1, 0
  ),
  nrow = 2,
  byrow = TRUE
)
e <- c(0, 0, 0, 1, 1, 1)

# Synthetic per-sample mutation counts.
nmut <- c(3, 4, 5, 6, 7, 8)

result <- mlr.v2(mut, e, nmut)
result$pvalue

optifix. Optimise with fixed parameters

Description

its like optim, but with fixed parameters.

Usage

optifix(
  par,
  fixed,
  fn,
  gr = NULL,
  ...,
  method = c("Nelder-Mead", "BFGS", "CG", "L-BFGS-B", "SANN"),
  lower = -Inf,
  upper = Inf,
  control = list(),
  hessian = FALSE
)

Arguments

Details

specify a second argument 'fixed', a vector of TRUE/FALSE values. If TRUE, the corresponding parameter in fn() is fixed. Otherwise its variable and optimised over.

The return thing is the return thing from optim() but with a couple of extra bits - a vector of all the parameters and a vector copy of the 'fixed' argument.

Written by Barry Rowlingson <b.rowlingson@lancaster.ac.uk> October 2011

This file released under a CC By-SA license: http://creativecommons.org/licenses/by-sa/3.0/

and must retain the text: "Originally written by Barry Rowlingson" in comments.

plot phenotype, mutation and annotation for a gene across samples

Description

plot phenotype, mutation and annotation for a gene across samples

Usage

plot_mut(
  gene_name,
  mut,
  pheno,
  totalnttype = 96,
  anno_dir = ".",
  output_prefix = "plot",
  output_dir = tempdir()
)

Arguments

Value

A numeric vector returned by the final barplot() call, giving the coordinates of the plotted bars. The function is primarily called for its side effect of drawing mutation and phenotype summary plots for the selected gene.

Examples

phenof <- system.file(
  "extdata", "example_phenotypes.txt",
  package = "diffdriver"
)
mutf <- system.file(
  "extdata", "example_mutations.txt",
  package = "diffdriver"
)

pheno <- read.table(phenof, header = TRUE)
mut <- read.table(mutf, header = TRUE)

## Not run: 
# Download the annotation files and replace this with their directory.
annodir <- "/path/to/annodir9"

plot_mut(
  gene_name = "PIK3CA",
  mut = mut,
  pheno = pheno,
  totalnttype = 9,
  anno_dir = annodir,
  output_dir = tempdir()
)

## End(Not run)