$computeMap | R Documentation |
Method for inferring parental phase (e.g., ordered parental genotype pair (OPGP)) and estimating recombination fractions in full-sib families.
BCobj$computeMap(chrom = NULL, init_r = 0.01, ep = 0.001, method = NULL, err = TRUE,
multiErr = FALSE, mapped = TRUE, nThreads = 1, inferOPGP = TRUE, rfthres = 0.1)
FSobj$computeMap(chrom = NULL, init_r = 0.01, ep = 0.001, method = NULL, sexSpec = FALSE, err = TRUE,
multiErr = FALSE, mapped = TRUE, nThreads = 1, inferOPGP = TRUE, rfthres = 0.1)
ICobj$computeMap(chrom = NULL, init_r = 0.01, ep = 0.001, method = "EM", err = TRUE,
multiErr = FALSE, mapped = TRUE, nThreads = 1, inferOPGP = TRUE, rfthres = 0.1)
init_r |
A numeric value giving the initial values for the recombination fractions. Each recombination fraction parameter is set to the same initial value. |
ep |
A numeric value giving the initial value for the sequencing error parameter. |
sexSpec |
Logical value. If |
multiErr |
Locical value. If |
method |
A character string specifying whether optimization should be performed using
direct maximization ( |
nThreads |
An integer value giving the number of threads to use in computing the likelihood in parallel. |
chrom |
A integer vector giving the indices of the chromosomes (or linkage groups) to be computed. |
err |
Locical value. If |
mapped |
Locial value. If |
inferOPGP |
Logical value. If |
rfthres |
Numeric value. Print output highlighting SNP pairs with high recombination fraction estimates. Helpfully for identifying potentially probmatic SNPs. |
This function infers the parental phase (or ordered parental genotype pair (OPGP)) and estimates adjacent recombination fractions using the hidden Markov model (HMM) approach as described in \insertCitebilton2018genetics1;textualGUSMap.
The optimization of the likelihood for the HMM is performed using either the Expectation-Maximumization (EM) algorithm
(method="EM"
), direct numeric optimization via the optim
function (method="optim"
), or using
20 iterations of the EM algorithm followed by optimization via direct numeric optimization (method=NULL
).
The likelihood computations (and computation of derivatives if required) are scaled using
forward and backward recursion to avoid overflow issues and are performed in C. These computations
are also parallelization via the OpenMP package, where the argument nThreads
specifies
how many threads to use. Be careful not to set nThreads
to more than the number of threads available
on your computer (or bad things will happen). In addition, if the package is complied without OpenMP, then this
parallelization has no effect and the likelihood is computed in serial.
If mapped = TRUE
, then combined linkage groups must have been formed from the $addBIsnps
function
first (and preferably ordered from the $orderLG
function).
Timothy P. Bilton and Chris Scott
bilton2018genetics1GUSMap
BC
, FS
, IC
#### Case 1: Compute linkage map from linkage groups
## Simulate some sequencing data
set.seed(6745)
config <- list(list(sample(c(1,2,4), size=30, replace=TRUE)))
F1data <- simFS(0.01, config=config, meanDepth=10, nInd=50, epsilon=0.005)
## Compute 2-point recombination fractions
F1data$rf_2pt(nClust=1)
## create and order linkage groups
F1data$createLG()
F1data$addBIsnps()
F1data$orderLG(ndim=5)
## Compute the linkage map
F1data$computeMap()
#### Case 2: Compute map using original assembly order
F1data$computeMap(mapped = FALSE)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.