Package 'RTIGER'

Title: HMM-Based Model for Genotyping and Cross-Over Identification
Description: Our method integrates information from all sequenced samples, thus avoiding loss of alleles due to low coverage. Moreover, it increases the statistical power to uncover sequencing or alignment errors <doi:10.1093/plphys/kiad191>.
Authors: Rafael Campos-Martin [cre] , Sophia Schmickler [aut], Manish Goel [ctb], Korbinian Schneeberger [aut], Achim Tresch [aut]
Maintainer: Rafael Campos-Martin <[email protected]>
License: GPL (>= 2)
Version: 2.1.0
Built: 2025-02-22 04:11:54 UTC
Source: https://github.com/rfael0cm/rtiger

Help Index

The autosome chromosome lengths for Arabidopsis Thaliana.

Description

The autosome chromosome lengths for Arabidopsis Thaliana.

Author(s)

Rafael Campos-Martin

Obtain number of Cross-Over events per sample and chromosome.

Description

Obtain number of Cross-Over events per sample and chromosome.

Usage

calcCOnumber(object)

Arguments

object

a RViterbi object.

Value

Matrix m x n. M number of samples and N chromosomes.

#' @return a matrix with n chromosomes and m samples (n x m) and the number of CO events.

Examples

data("fittedExample")
co.num = calcCOnumber(myDat)

Function to developers. It runs one EM step

Description

Function to developers. It runs one EM step

Usage

dev(psi, rigidity = NULL, nstates = 3, transition = NULL, start = NULL)

Arguments

psi

list of psi probabilities.
rigidity

Rigidity value.
nstates

Number of states.
transition

transition matrix
start

initial probabilities

Value

List with updates probabilites

Call Julia code to fit the values

Description

Call Julia code to fit the values

Usage

fit(rtigerobj, max.iter , eps,
trace, all = TRUE, random = FALSE,
specific = FALSE, nsamples = 20,
post.processing = TRUE)

Arguments

rtigerobj

an RTIGER object.
max.iter

maximum number of iterations to acomplish by the EM.
eps

differnece threshold to halt the EM.
trace

logical value whether to trace the changes in the parameters along the iterations.
all

logical value whether to use all data to fit the model.
random

if all FALSE use random samples.
specific

if all FALSE use specific samples.
nsamples

if random TRUE, how many samples to use.
post.processing

logical value, whether to run post.processing process.

Value

RTIGER object

Examples

## Not run: 
data("fittedExample")
sourceJulia()
myfit = fit(myDat, max.iter = 2, eps=0.01,
            trace = TRUE, all = TRUE,
            random = FALSE, specific = FALSE,
            nsamples = 20, post.processing = TRUE)


## End(Not run)

Load data

Description

Load data

Usage

generateObject(experimentDesign = NULL,nstates = 3, rigidity=NULL,
seqlengths = NULL, verbose = TRUE)

Arguments

experimentDesign

a data Frame that contains minimum a column with the files direction (name of the column files) and another with a shorter name to be used inside the function.
nstates

the number of states to be fitted in the model. A standard setting would use 3 states (Homozygous1, Heterozygous, and Homozygous2).
rigidity

an integer number specifying the rigidity parameter to be used.
seqlengths

a named vector with the chromosome lenghts of the organism that the user is working with.
verbose

logical value. Whether to print info messages.

Value

RTIGER object

Examples

data("ATseqlengths")
path = system.file("extdata",  package = "RTIGER")
files = list.files(path, full.names = TRUE)
nam = sapply(list.files(path ), function(x) unlist(strsplit(x, split = "[.]"))[1])
expDesign = data.frame(files = files, name = nam)
names(ATseqlengths) = paste0("Chr", 1:5)
myres = generateObject(experimentDesign = expDesign,
              seqlengths = ATseqlengths,
              rigidity = 10
)

A fitted example using three own samples of Arabidopsis. More information in publication:

Description

A fitted example using three own samples of Arabidopsis. More information in publication:

Author(s)

Rafael Campos-Martin

Find the otimum R value for a given data set

Description

Find the otimum R value for a given data set

Usage

optimize_R(object,
max_rigidity = 2^9, average_coverage = NULL, crossovers_per_megabase = NULL,
save_it = FALSE, savedir = NULL)

Arguments

object

an RTIGER object
max_rigidity

R values will be explored up the value given in this parameter. Default = 2^9
average_coverage

For conservative results set it to the lowest average coverage of a sample in your experiment, or evne to the lowest average coverage in a (sufficiently large) region in one of your samples. The lower the value, the more conservative (higher) our estimates of the false positive segments rates. If it is not provided it will be computed as the average of all data points.
crossovers_per_megabase

For conservative results set it to the highest ratio of a sample in your experiment. The higher the value, the more conservative (higher) our estimates of the false positive segments rates. If it is not provided it will be computed as the average of all samples.
save_it

logical values if the results should be saved. Plots might be complicated to interpret. We suggest to read the manuscript to understand them (https://doi.org/10.1093/plphys/kiad191)
savedir

if results are saved, in which directory.

Value

A value with the optimum rigidity for the data set.

Examples

data("fittedExample")
bestR = optimize_R(myDat)

Obtain number of Cross-Over events per sample and chromosome.

Description

Obtain number of Cross-Over events per sample and chromosome.

Usage

plotCOs(object, file = NULL)

Arguments

object

a RViterbi object.
file

file where to save the plot for CO numbers

Value

a plot

Examples

data("fittedExample")
co.num = calcCOnumber(myDat)

Load, Fit, and plot

Description

Load, Fit, and plot

Usage

RTIGER(expDesign, rigidity=NULL, outputdir=NULL, nstates = 3,
seqlengths = NULL, eps=0.01, max.iter=50, autotune = FALSE,
max_rigidity = 2^9, average_coverage = NULL,
crossovers_per_megabase = NULL, trace = FALSE,
tiles = 4e5, all = TRUE, random = FALSE, specific = FALSE,
nsamples = 20, post.processing = TRUE, save.results = TRUE, verbose = TRUE)

Arguments

expDesign

a data Frame that contains minimum a column with the files direction (name of the column files) and another with a shorter name to be used inside the function.
rigidity

an integer number specifying the rigidity parameter to be used.
outputdir

a character string that specifies the directory in which to save the results form the function.
nstates

the number of states to be fitted in the model. A standard setting would use 3 states (Homozygous1, Heterozygous, and Homozygous2).
seqlengths

a named vector with the chromosome lenghts of the organism that the user is working with.
eps

the threshold of the difference between the parameters value between the previous and actuay iteration to stope de EM algorithm.
max.iter

maximum number of iterations of the EM algorithm before to stop in case that eps has not been achieved.
autotune

Logical value if the R-value should be tuned by our algorithm. This will take longer as it needs a first training with the rigidity value provided by the user and then the optimization step is carried. Finally, a training using the optimum R will be performed and results for the optimum R will be returned.
max_rigidity

If autotune true, R values will be explored up the value given in this parameter. Default = 2^9
average_coverage

If autotune true, for conservative results set it to the lowest average coverage of a sample in your experiment, or evne to the lowest average coverage in a (sufficiently large) region in one of your samples. The lower the value, the more conservative (higher) our estimates of the false positive segments rates. If it is not provided it will be computed as the average of all data points.
crossovers_per_megabase

If autotune true, for conservative results set it to the highest ratio of a sample in your experiment. The higher the value, the more conservative (higher) our estimates of the false positive segments rates. If it is not provided it will be computed as the average of all samples.
trace

logical value. Whether or not to keep track of the parameters for the HMM along the iterations. Deafault FALSE
tiles

length of the tiles by which the genome will be segmented in order to compute the ratio of COs in the complete dataset.
all

logical value. Whether to use the complete data set to fit the rHMM. default TRUE.
random

Logical value. Choose randomly a subset of the complete dataset to fit the rHMM. Default FALSE
specific

Logical value to specify which samples to take.
nsamples

if random TRUE, how many samples should be taken randomly.
post.processing

Logical value. Whether to run an extra step that fine maps the segment borthers. Default TRUE
save.results

Logical value, whether to generate and save the plots and igv files.
verbose

Logical, whether to print info to console.

Value

Matrix m x n. M number of samples and N chromosomes.

RTIGER object

Examples

## Not run: 
data("ATseqlengths")
sourceJulia()
path = system.file("extdata",  package = "RTIGER")
files = list.files(path, full.names = TRUE)
nam = sapply(list.files(path ), function(x) unlist(strsplit(x, split = "[.]"))[1])
expDesign = data.frame(files = files, name = nam)
names(ATseqlengths) = paste0("Chr", 1:5)
myres = RTIGER(expDesign = expDesign,
               outputdir = "/home/campos/Documents/outputjulia/",
               seqlengths = ATseqlengths,
               rigidity = 4,
               max.iter = 2,
               trace = FALSE,
               save.results = TRUE)

## End(Not run)

This class is a generic container for RTIGER analysis

Description

This class is a generic container for RTIGER analysis

Slots

matobs

Nested lists. the first level is a list of samples. For each sample there are 5 matrices that contains the allele counts for each position.

params

a list with the parameters after training.

info

List with phenotipic data of the samples.

Viterbi

List of chromosomes with the viterbi path per sample.

Probabilities

Computed probabilites for the EM algorithm.

num.iter

Number of iterations needed to stop the EM algorithm.

Installs the needed packages in JULIA to run the EM algorithm for rHMM.

Description

Installs the needed packages in JULIA to run the EM algorithm for rHMM.

Usage

setupJulia(JULIA_HOME = NULL)

Arguments

JULIA_HOME

the file folder which contains julia binary, if not set, JuliaCall will look at the global option JULIA_HOME, if the global option is not set, JuliaCall will then look at the environmental variable JULIA_HOME, if still not found, JuliaCall will try to use the julia in path.

Value

empty

Function needed before using RTIGER() function. It loads the scripts in Julia that fit the rHMM.

Description

Function needed before using RTIGER() function. It loads the scripts in Julia that fit the rHMM.

Usage

sourceJulia()

Value

empty