sigminer (version 0.1.11)

get_components: Fit optimal number of mixture model components

Description

Apply mixture modelling to breakdown each feature distribution into mixtures of Gaussian or mixtures of Poison distributions using flexmix package. The order of features is 'Segment size', 'Breakpoint count per 10 Mb', 'Length of oscillating copy-number chain', 'Breakpoint count per arm', 'Copy number change', 'Absolute copy number'.

Usage

get_components(CN_features, seed = 123456, min_comp = 2,
  max_comp = 10, min_prior = 0.001, model_selection = "BIC",
  nrep = 1, niter = 1000)

Arguments

CN_features

a list generate from get_features() function.

seed

seed number.

min_comp

minimal number of components to fit, default is 2. Can also be a vector with length 6, which apply to each feature.

max_comp

maximal number of components to fit, default is 10. Can also be a vector with length 6, which apply to each feature.

min_prior

minimal prior value, default is 0.001. Details about custom setting please refer to flexmix package.

model_selection

model selection strategy, default is 'BIC'. Details about custom setting please refer to flexmix package.

nrep

number of run times for each value of component, keep only the solution with maximum likelihood.

niter

maximal number of iteration to achive converge.

Value

a list contain flexmix object of copy-number features.

See Also

Other internal calculation function series: get_ArmLocation, get_LengthFraction, get_cnlist, get_cnsummary_sample, get_features, get_matrix

Examples

# NOT RUN {
# Load copy number features
load(system.file("extdata", "toy_cn_features.RData",
  package = "sigminer", mustWork = TRUE
))
cn_components <- get_components(cn_features)
# }