protected.mixML

Attempts to catch numerical errors in maximum likelihood computation, by replacing logL values by a default value instead of NaN or other failures

Methods for inference about/under complex relationships using
peak height data from DNA mixtures: the most basic example would be testing whether a contributor to
a mixture is the father of a child of known genotype. This provides most of the functionality
of the 'KinMix' package, but with some loss of efficiency and restriction on problem size,
as the latter uses 'RHugin' as the Bayes net engine, while this package uses 'gRain'.
The package implements the methods introduced in
Green, P. J. and Mortera, J. (2017) <doi:10.1016/j.fsigen.2017.02.001> and
Green, P. J. and Mortera, J. (2021) <doi:10.1111/rssc.12498>.

Peter Green

KinMixLite

Inference About Relationships from DNA Mixtures

protected.mixML function

<dl> <dt>mixture</dt>
<dd>A DNAmixture object.</dd> <dt>pars</dt>
<dd>A mixpar parameter used as a starting value for the optimisation.</dd> <dt>constraints</dt>
<dd>Equality constraint function as function of an array of parameters.</dd> <dt>phi.eq</dt>
<dd>Should the mixture proportions be the same for all traces? Defaults to FALSE.</dd> <dt>val</dt>
<dd>Vector of values to be satisfied for the equality constraints.</dd> <dt>trace</dt>
<dd>Print the evaluations of the likelihood-function during optimisation?</dd> <dt>order.unknowns</dt>
<dd>Should unknown contributors be ordered according to decreasing contributions? Defaults to TRUE.</dd> <dt>...</dt>
<dd>Further arguments to be passed on to solnp.</dd> <dt>default</dt>
<dd>value of logL to be used if numerical errors are encountered</dd></dl>

Arguments

Peter Green (P.J.Green@bristol.ac.uk)

Author

Protect against numerical errors in maximum likelihood computation — protected.mixML

<dl>

 <dt>mixture</dt>
<dd>A DNAmixture object.</dd>

 <dt>pars</dt>
<dd>A mixpar parameter used as a starting value for the optimisation.</dd>

 <dt>constraints</dt>
<dd>Equality constraint function as function of an array of parameters.</dd>

 <dt>phi.eq</dt>
<dd>Should the mixture proportions be the same for all traces? Defaults to FALSE.</dd>

 <dt>val</dt>
<dd>Vector of values to be satisfied for the equality constraints.</dd>

 <dt>trace</dt>
<dd>Print the evaluations of the likelihood-function during optimisation?</dd>

 <dt>order.unknowns</dt>
<dd>Should unknown contributors be ordered according to decreasing contributions? Defaults to TRUE.</dd>

 <dt>...</dt>
<dd>Further arguments to be passed on to solnp.</dd>

 <dt>default</dt>
<dd>value of logL to be used if numerical errors are encountered</dd>

</dl>

protected.mixML: Protect against numerical errors in maximum likelihood computation

Description

Usage

Value

Arguments

Author