contLikINT

Share:

Description

contLikINT marginalizes the likelihood of the STR DNA mixture given some assumed a bayesian model by numerical integration.

Usage

1
2
3
4
contLikINT(nC, samples, popFreq, lower, upper, refData = NULL,
  condOrder = NULL, knownRef = NULL, xi = NULL, prC = 0,
  reltol = 0.01, threshT = 50, fst = 0, lambda = 0, pXi = function(x)
  1, kit = NULL, scale = 0, maxEval = 0)

Arguments

nC

Number of contributors in model.

samples

A List with samples which for each samples has locus-list elements with list elements adata and hdata. 'adata' is a qualitative (allele) data vector and 'hdata' is a quantitative (peak heights) data vector.

popFreq

A list of allele frequencies for a given population.

lower

Lower bounds of parameters. Must be in following order: mx1,..,mx_(nC-1),mu,sigma,beta,xi.

upper

Upper bounds of parameters. Must be in following order: mx1,..,mx_(nC-1),mu,sigma,beta,xi.

refData

Reference objects has locus-list element [[i]] with a list element 'r' which contains a 2 long vector with alleles for each references.

condOrder

Specify conditioning references from refData (must be consistent order). For instance condOrder=(0,2,1,0) means that we restrict the model such that Ref2 and Ref3 are respectively conditioned as 2. contributor and 1. contributor in the model.

knownRef

Specify known non-contributing references from refData (index). For instance knownRef=(1,2) means that reference 1 and 2 is known non-contributor in the hypothesis. This affectes coancestry correction.

xi

A numeric giving stutter-ratio if it is known. Default is NULL, meaning it is integrated out.

prC

A numeric for allele drop-in probability. Default is 0.

reltol

Required relative tolerance error of evaluations in integration routine. Default is 0.001.

threshT

The detection threshold given. Used when considering probability of allele drop-outs.

fst

is the coancestry coeffecient. Default is 0.

lambda

Parameter in modeled peak height shifted exponential model. Default is 0.

pXi

Prior function for xi-parameter (stutter). Flat prior on [0,1] is default.

kit

shortname of kit: "ESX17","ESI17","ESI17Fast","ESX17Fast","Y23","Identifiler","NGM","ESSPlex","ESSplexSE","NGMSElect","SGMPlus","ESX16", "Fusion","GlobalFiler"

scale

used to make integrale calculateable for small numbers. For scale!=0, integrale must be scaled afterwards with exp(-scale) to be correct.

maxEval

Maximum number of evaluations in the adaptIntegrate function. Default is 0 which gives an infinite limit.

Details

The procedure are doing numerical integration to approximate the marginal probability over the noisance parameters. Mixture proportions have flat prior.

Function calls procedure in c++ by using the package Armadillo and Boost.

Value

ret A list(margL,deviation,nEvals) where margL is Marginalized likelihood for hypothesis (model) given observed evidence, deviation is the confidence-interval of margL, nEvals is number of evaluations.

Author(s)

Oyvind Bleka <Oyvind.Bleka.at.fhi.no>

References

Hahn,T. (2005). CUBA - a library for multidimensional numerical integration. Computer Physics Communications, 168(2),78-95.