[adegenet-commits] r875 - pkg/man

[noreply at r-forge.r-project.org]

Author: jombart
Date: 2011-05-24 12:11:31 +0200 (Tue, 24 May 2011)
New Revision: 875

Modified:
   pkg/man/glSim.Rd
Log:
completed the doc of glSim.
shitfuck: multicore is gone.


Modified: pkg/man/glSim.Rd
===================================================================
--- pkg/man/glSim.Rd	2011-05-24 09:00:42 UTC (rev 874)
+++ pkg/man/glSim.Rd	2011-05-24 10:11:31 UTC (rev 875)
@@ -4,13 +4,12 @@
 \title{Simulation of simple genlight objects}
 \description{
   The function \code{glSim} simulates simple SNP data with the
-  possibility of contrasted structures between two groups. This can be
-  useful to test methods for defining characteristic alleles. Returned
+  possibility of contrasted structures between two groups. Returned
   objects are instances of the class \linkS4class{genlight}.
 }
 \usage{
 glSim(n.ind, n.snp.nonstruc, n.snp.struc = 0, grp.size = round(n.ind/2), 
-    ploidy = 1, alpha = 0, block.size = NULL)
+    ploidy = 1, alpha = 0, block.size = NULL, LD = FALSE)
 }
 \arguments{
   \item{n.ind}{an integer indicating the number of individuals to be simulated.}
@@ -27,25 +26,54 @@
     size.}
   \item{ploidy}{an integer indicating the ploidy of the simulated
     genotypes.}
-  \item{alpha}{asymmetry parameter: a numeric value between 0 and 0.5, used to enforce
-    allelic differences between the groups (see details).}
+  \item{alpha}{asymmetry parameter: a numeric value between 0 and 0.5,
+    used to enforce allelic differences between the groups (see
+    details); ignored if \code{LD=TRUE}.}
   \item{block.size}{an optional integer indicating the number of SNPs to
-    be handled at a time during the simulations. By default, all SNPs are
-    simulated at the same time, but RAM can limit this operation. Using
-    blocks of a few hundred or thousand SNPs decreases RAM requirement at
-    a cost of more computational time.}
-
+    be handled at a time during the simulations. By default, all SNPs
+    are simulated at the same time, but RAM can limit this
+    operation. Using blocks of a few hundred or thousand SNPs decreases
+    RAM requirement at a cost of more computational time. When
+    \code{LD=TRUE}, large blocks will come at a large costs in terms of
+    computational time and RAM, since the underlying matrices of
+    correlation will be large.}
+  \item{LD}{a logical indicating whether loci should be displaying
+    linkage disequilibrium (TRUE) or be generated independently
+    (FALSE, default). When set to TRUE, data are generated by blocks of
+    correlated SNPs (see details).}
 }
 \details{
   === Allele frequencies in contrasted groups ===
+  
   When \code{n.snp.struc} is greater than 0, some SNPs are simulated in
   order to differ between groups (noted 'A' and 'B'). Such differences
-  are achieved by using different frequencies of the second allele for A
-  and B, denoted \eqn{p_A} and \eqn{p_B}. For a given SNP, \eqn{p_A} is
-  drawn from a uniform distribution between 0 and (0.5 -
-  alpha). \eqn{p_B} is then computed as 1 - \eqn{p_A}. Therefore,
-  differences between groups are mild for alpha=0, and total for alpha =
-  0.5.
+  can be achieved differently depending on whether loci are independent
+  (\code{LD=FALSE}), or not (\code{LD=TRUE}). In the first case,
+  different patterns between groups are achieved by using different
+  frequencies of the second allele for A and B, denoted \eqn{p_A} and
+  \eqn{p_B}. For a given SNP, \eqn{p_A} is drawn from a uniform
+  distribution between 0 and (0.5 - alpha). \eqn{p_B} is then computed
+  as 1 - \eqn{p_A}. Therefore, differences between groups are mild for
+  alpha=0, and total for alpha = 0.5.
+
+  Whenever loci are linked (\code{LD=TRUE}), this option is no longer
+  available. Differences between groups merely occur by drawing alleles
+  from randomly generated, group-specific allele frequencies.
+
+  
+  === Linked or independent loci ===
+  
+  Independent loci (\code{LD=FALSE}) are simulated using the standard
+  binomial distribution, with randomly generated allele
+  frequencies. Linked loci (\code{LD=FALSE}) are trickier towe need to
+  simulate discrete variables with pre-defined correlation structure.
+
+  Here, we first generate deviates from multivariate normal
+  distributions with randomly generated correlation structures. These
+  variables are then discretized using the quantiles of the
+  distribution. Further improvement of the procedure will aim at i)
+  specifying the strength of the correlations between blocks of alleles
+  and ii) enforce contrasted structures between groups.
 }
 \value{
   A \linkS4class{genlight} object.
@@ -77,5 +105,9 @@
 x <- glSim(100, 1e3, n.snp.struc=100, ploid=2, alpha=0.4)
 plot(x)
 
+##  same thing, loci with LD structures
+x <- glSim(100, 1, n.snp.struc=100, ploid=2, alpha=0.4, LD=TRUE, block.size=100)
+plot(x)
+
 }
 \keyword{multivariate}
\ No newline at end of file