False Discovery Rate using a Mixture Model

This menu can be used to fit a mixture model to a distribution of probabilities. The two components of the mixture can be thought as those probes which are showing differential expression (the Beta/Gamma component) and those not responding (the Uniform component) whose probabilities values then form a random sample from the null distribution of the test statistic.

The false discovery rate (FDR), false rejection rate (FRR) and power of a test by modelling significance values as a 2-component mixture of Uniform and Beta or Gamma densities, Allison et al. (2002) are estimated. The context is multiple testing, with data from any situation where the same simple hypothesis, H_o, is tested many times, such as in transcriptomics (microarrays), metabolomics and proteomics. These tests generate a large number of significance values which, under H₀, have a Uniform distribution and, under H_a, can be modelled as a Beta or truncated Gamma density. The parameters of this mixture distribution to derive the False Discovery Rate, Prob(H₀/D_a), the False Rejection Rate, Prob(H_a/D₀) and the Power of the test, Prob(D_a/H_a), each as a function of p_crit are estimated. Here D_a denotes the event "p<pcrit".

The mixture model parameterization takes a proportion P from the Uniform distribution, and (1 - P) from either a Beta or Gamma distribution. The Gamma parameterization is:

f(x) = (1/b)^A / Gamma(A) × exp(-x/B) × x^(A-1), truncated at x=1,

and the Beta parameterization is:

f(x) = x^(A-1) × (1-x)^(B-1) / Beta(A; B).

A range of graphs are plotted after the mixture model has been fitted. Initial values for the 3 parameters P,A and B need to be provided.

Available Data

This lists data structures appropriate for the edit box which currently has focus. You can double-click a name to enter it in the edit box.

Distribution

The distribution for the responding probes to mix with the uniform distribution, either:

Beta	The Beta distribution
Truncated Gamma	The Gamma distribution truncated at 1.

These distributions are described above.

Probabilities to Model

This variate contains the probabilities to which the mixture distribution is fitted. All the values in the variate should be between 0 and 1 (although missing values (*) are allowed).

Initial Mixing Proportion

The initial value for proportion of the sample which are from the Uniform distribution, i.e. those probes which are not showing any differential expression, but whose effects are caused by random measurement errors. The iterations will start at this value and try and find optimal parameters. If this is a long way from the optima, then the optimization iterations may not converge.

Initial A value

The initial value of A parameter for the Beta or Truncated Gamma distribution. The iterations will start at this value and try and find optimal parameters. If this is a long way from the optima, then the optimization iterations may not converge.

Initial B value

The initial B parameter for the Beta or Truncated Gamma distribution. The iterations will start at this value and try and find optimal parameters. If this is a long way from the optima, then the optimization iterations may not converge.

Action Buttons

Run	Run the analysis.
Cancel	Close the menu without further changes.
Options	Opens a dialog where additional options and settings can be specified for the analysis.
Defaults	Set the menu settings back to the default settings. Clicking the right mouse on this button produces a pop-up menu where you can choose to set the menu using the currently stored defaults or the GenStat default settings.
Store	Opens a dialog to specify names of structures to store the results from the analysis. The names to save the structures should be supplied before running the analysis.