modele: Create a compositional model for simulations
In SARP.compo: Network-Based Interpretation of Changes in Compositional Data

Description Usage Arguments Details Value Author(s) See Also Examples

These functions create and plot a model of compositionnal data for two or more conditions.

modele_compo( medianes, en.log = FALSE,
              noms = colnames( medianes ),
              conditions = rownames( medianes ),
              reference = NULL, total = 1 )

## S3 method for class 'SARPcompo.modele'
plot( x,
     xlab = "Composant",
     ylab.absolu = "Quantit\u00e9", ylab.relatif = "Fraction",
     taille.noeud = 50, ... )

`medianes`	A matrix giving the medians of all components quantities in each condition. Each row of this matrix corresponds to a different condition; each column, to one of the components.
`en.log`	If `TRUE`, the values in the matrix are given in the log scale.
`noms`	Names of the components. If provided, should be a character vector whose length is equal to the number of columns of `medianes`.
`conditions`	Names of the different conditions. If provided, should be a character vector whose length is equal to the number of rows of `medianes`.
`reference`	A character vector giving the names of the components used as reference (typically, reference genes in qRT-PCR).
`total`	The total amount. The sum of amounts in each condition will equal this total, when the data are made compositionnal.

Arguments for the plot method

`x`	The modele to be plotted
`xlab`	Legend for the X axis
`ylab.absolu`	Legend for the Y axis, in the amount scale (no constrain)
`ylab.relatif`	Legend for the Y axis, in the compositional scale.
`taille.noeud`	The plot size of nodes of the theoretical graph
`...`	Additionnal parameters for `plot`, which is used internally.

The modele_compo function creates a compositionnal data model using the quantites provided: it converts amounts in fractions of the total amount for each condition, then computes the theoretical graph showing classes of equivalents components, that is components that have the same evolution between the two conditions. If more than two conditions are given, graphs correspond to comparison of each condition with the first one.

The plot methods represents the original quantities, the quantities after conversion in compositional data ant the theoretical graph.

An object of class SARPcompo.modele, with a plot method. It is a list with the following elements:

`Absolue`	The matrix of quantities in amount scale
`Relative`	The matrix of quantities in compositional data scale
`Graphes`	A list of length `nrow(medianes) - 1`. Each element of the list gives, for the corresponding condition, the matrix of all ratios of pairwise ratios between condition and the first condition (element `M.rapports`), the corresponding connectivity matrix (element `M.connexion`), the graph of component changes compared to the first condition (element `Graphe`, an igraph object) and the list of components of this graph (element `Connexe`, obtained from the `components` function.

It also stores a few informations as attributes.

Emmanuel Curis (emmanuel.curis@parisdescartes.fr)

estimer.puissance and estimer.alpha to use these models in simulations to study power and type I error of the method in a given situation.

  ## Create a toy example: four components, two conditions
  ##  components 1 and 2 do not change between conditions
  ##  component  3 is doubled
  ##  component  4 is halfed
  me <- rbind( 'A' = c( 1, 1, 1, 1 ),
               'B' = c( 1, 1, 2, 0.5 ) )
  colnames( me ) <- paste0( "C-", 1:4 )

  md <- modele_compo( me )

  ## Plot it...
  plot( md )

  ## What is approximately the power to detect that something changes
  ## between conditions A and B using a Student test
  ## with a CV of around 50 % ?
  ##  (only a few simulations for speed, should be increased )
  puissance <- estimer.puissance( md, cv = 0.50, B = 50, f.p = student.fpc )
  plot( puissance )