orderCross: Order markers Order markers within linkage groups using...

Description Usage Arguments Value

View source: R/orderCross.R

Description

This function orders markers within linkage groups using a simulated annealing heuristic. The underlying implementation is a C++ reimplementation of the fortran code arsa.f from the seriation package. The reimplementation allows for multithreading, and is therefore much faster. It also fixes a couple of bugs in the original code.

Parameters cool and tmin are standard simulated annealing parameters, and decreasing cool increases the amount of computation effort. Parameter nReps gives the number of independent replications of the simulated annealing algorithm to be used. The result of the best replication is then chosen.

Parameter maxMove gives the maximum number of positions by which to shift a marker, as part of a step within the simulated annealing algorithm. The computational effort of determining whether a proposed move of a particular marker should be accepted, depends on the number of positions by which it is moved. So if the ordering is already approximately correct at the start of the algorithm, proposals that move markers by large distances are expensive, and also unneccessary. These types of proposed changes to the ordering can be avoided by setting maxMove to some positive value, maybe one tenth of the number of markers.

Parameter effortMultiplier simply increases or decreases the amount of computational effort. A value of 0.5 requires half as much effort, a value of 1.0 uses the default amount of effort, and a value of 2.0 requires twice as much computational effort.

Parameter randomStart controls the starting point of each replication of the algorithm. If this parameter is TRUE, then every replication starts form an independent random ordering. If this parameter is FALSE, then every replication starts from the marker ordering given in the input object.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
orderCross(
  mpcrossLG,
  cool = 0.5,
  tmin = 0.1,
  nReps = 1,
  maxMove = 0,
  effortMultiplier = 1,
  randomStart = TRUE,
  verbose = FALSE
)

Arguments

mpcrossLG

An object of class mpcrossLG, containing genetic data and linkage groups.

cool

Rate of cooling

tmin

Minimum temperature

nReps

Number of independent replications of the simulated annealing algorithm

maxMove

Maximum number of positions by which to shift a single marker, as part of the simulated annealing. A value of zero indicates no limit.

effortMultiplier

Multiplier for the amount of computational effort

randomStart

If TRUE, start from the current ordering

verbose

If TRUE, generate more detailed output

Value

An object of class mpcrossLG, identical to the input except with the markers rearranged.


mpMap2 documentation built on Sept. 13, 2020, 5:17 p.m.