Skip to contents

Calculate Marginal Likelihood for a Single Segment

Source: R/Modeling.R
CalSegmentLikelihood.Rd

Computes the marginal likelihood for a segment given copy number, MAF, purity, and model parameters.

Usage

CalSegmentLikelihood(C_i, B_i, mu, rho, sigma_C, k, lambda, gamma, epsilon)

Arguments

C_i

Numeric. Integer copy number for the segment.

B_i

Numeric. Observed B-allele frequency (MAF) for the segment.

mu

Numeric. Diploid coverage scale factor (e.g. A value of 1 indicates that the segment mean from GATK does not require adjustment. Additionally, the pseudo-diploid coverage is set to 100.).

rho

Numeric. Tumor purity (fraction between 0 and 1).

sigma_C

Numeric. Not directly used in this function, but included for compatibility.

k

Numeric. Beta distribution concentration parameter.

lambda

Numeric. Exponential decay parameter for the prior.

gamma

Numeric. Weight for the prior in the likelihood calculation.

epsilon

Numeric. Small value to avoid log(0) and zero parameters in beta.

Value

A data frame with columns: major, minor, CN, ccf, Bio_diff, prior, expected_MAF, MAF_ll, weighted_prior, exp_MAF_ll, exp_prior, MAF_likelihood, Segcov, MAF, mu, rho.

Details

The returned data frame contains the following columns:

  • major: Major allele count

  • minor: Minor allele count

  • CN: Total copy number

  • ccf: Cancer cell fraction

  • Bio_diff: Biological difference

  • prior: Prior probability

  • expected_MAF: Expected MAF

  • MAF_ll: MAF log-likelihood

  • weighted_prior: Weighted prior

  • exp_MAF_ll: Expected MAF log-likelihood

  • exp_prior: Expected prior

  • MAF_likelihood: MAF likelihood

  • Segcov: Segment coverage

  • MAF: Observed MAF

  • mu: Mutation multiplicity

  • rho: Tumor purity