est_rf_hmm: Multipoint analysis using Hidden Markov Models in autopolyploids

Description

Performs the multipoint analysis proposed by Mollinari and Garcia (2019) in a sequence of markers

Usage

est_rf_hmm(
  input.seq,
  input.ph = NULL,
  thres = 0.5,
  twopt = NULL,
  verbose = FALSE,
  tol = 1e-04,
  est.given.0.rf = FALSE,
  reestimate.single.ph.configuration = TRUE,
  high.prec = TRUE
)
# S3 method for mappoly.map
print(x, detailed = FALSE, ...)
# S3 method for mappoly.map
plot(
  x,
  left.lim = 0,
  right.lim = Inf,
  phase = TRUE,
  mrk.names = FALSE,
  cex = 1,
  config = "best",
  P = "Parent 1",
  Q = "Parent 2",
  xlim = NULL,
  ...
)

Value

A list of class mappoly.map with two elements:

i) info: a list containing information about the map, regardless of the linkage phase configuration:

ploidy: the ploidy level
n.mrk: number of markers
seq.num: a vector containing the (ordered) indices of markers in the map, according to the input file
mrk.names: the names of markers in the map
seq.dose.p1: a vector containing the dosage in parent 1 for all markers in the map
seq.dose.p2: a vector containing the dosage in parent 2 for all markers in the map
chrom: a vector indicating the sequence (usually chromosome) each marker belongs as informed in the input file. If not available, chrom = NULL
genome.pos: physical position (usually in megabase) of the markers into the sequence
seq.ref: reference base used for each marker (i.e. A, T, C, G). If not available, seq.ref = NULL
seq.alt: alternative base used for each marker (i.e. A, T, C, G). If not available, seq.ref = NULL
chisq.pval: a vector containing p-values of the chi-squared test of Mendelian segregation for all markers in the map
data.name: name of the dataset of class mappoly.data
ph.thres: the LOD threshold used to define the linkage phase configurations to test

ii) a list of maps with possible linkage phase configuration. Each map in the list is also a list containing

seq.num: a vector containing the (ordered) indices of markers in the map, according to the input file
seq.rf: a vector of size (n.mrk - 1) containing a sequence of recombination fraction between the adjacent markers in the map
seq.ph: linkage phase configuration for all markers in both parents
loglike: the hmm-based multipoint likelihood

Arguments

input.seq: an object of class mappoly.sequence
input.ph: an object of class two.pts.linkage.phases. If not available (default = NULL), it will be computed
thres: LOD Score threshold used to determine if the linkage phases compared via two-point analysis should be considered. Smaller values will result in smaller number of linkage phase configurations to be evaluated by the multipoint algorithm.
twopt: an object of class mappoly.twopt containing two-point information
verbose: if TRUE, current progress is shown; if FALSE (default), no output is produced
tol: the desired accuracy (default = 1e-04)
est.given.0.rf: logical. If TRUE returns a map forcing all recombination fractions equals to 0 (1e-5, for internal use only. Default = FALSE)
reestimate.single.ph.configuration: logical. If TRUE returns a map without re-estimating the map parameters for cases where there is only one possible linkage phase configuration. This argument is intended to be used in a sequential map construction
high.prec: logical. If TRUE (default) uses high precision long double numbers in the HMM procedure
x: an object of the class mappoly.map
detailed: logical. if TRUE, prints the linkage phase configuration and the marker position for all maps. If FALSE (default), prints a map summary
...: currently ignored
left.lim: the left limit of the plot (in cM, default = 0).
right.lim: the right limit of the plot (in cM, default = Inf, i.e., will print the entire map)
phase: logical. If TRUE (default) plots the phase configuration for both parents
mrk.names: if TRUE, marker names are displayed (default = FALSE)
cex: The magnification to be used for marker names
config: should be 'best' or the position of the configuration to be plotted. If 'best', plot the configuration with the highest likelihood
P: a string containing the name of parent P
Q: a string containing the name of parent Q
xlim: range of the x-axis. If xlim = NULL (default) it uses the map range.

Author

Marcelo Mollinari, mmollin@ncsu.edu

Details

This function first enumerates a set of linkage phase configurations based on two-point recombination fraction information using a threshold provided by the user (argument thresh). After that, for each configuration, it reconstructs the genetic map using the HMM approach described in Mollinari and Garcia (2019). As result, it returns the multipoint likelihood for each configuration in form of LOD Score comparing each configuration to the most likely one. It is recommended to use a small number of markers (e.g. 50 markers for hexaploids) since the possible linkage phase combinations bounded only by the two-point information can be huge. Also, it can be quite sensible to small changes in 'thresh'. For a large number of markers, please see est_rf_hmm_sequential.

References

Mollinari, M., and Garcia, A. A. F. (2019) Linkage analysis and haplotype phasing in experimental autopolyploid populations with high ploidy level using hidden Markov models, _G3: Genes, Genomes, Genetics_. https://doi.org/10.1534/g3.119.400378

Examples

Run this code

    mrk.subset <- make_seq_mappoly(hexafake, 1:10)
    red.mrk <- elim_redundant(mrk.subset)
    unique.mrks <- make_seq_mappoly(red.mrk)
    subset.pairs <- est_pairwise_rf(input.seq = unique.mrks,
                                  ncpus = 1,
                                  verbose = TRUE)

    ## Estimating subset map with a low tolerance for the E.M. procedure
    ## for CRAN testing purposes
    subset.map <- est_rf_hmm(input.seq = unique.mrks,
                             thres = 2,
                             twopt = subset.pairs,
                             verbose = TRUE,
                             tol = 0.1,
                             est.given.0.rf = FALSE)
    subset.map
    ## linkage phase configuration with highest likelihood
    plot(subset.map, mrk.names = TRUE, config = "best")
    ## the second one
    plot(subset.map, mrk.names = TRUE, config = 2)

Run the code above in your browser using DataLab