sojournAnno: Estimate sojourn distribution parameters from posterior...

In biomvRCNS: Copy Number study and Segmentation for multivariate biological data

Description

Using prior information from previous studies or annotation data to determine sojourn distribution parameters

Usage

sojournAnno(xAnno, soj.type = "gamma", pbdist = NULL)

Arguments

xAnno	a GRanges / GRangesList / TxDb object, with its first meta column to represent the possible type of the range; Or a list object with named initial value vectors matching required parameters for a specific soj.type
soj.type	type of the sojourn distribution, following types are supported: 'gamma', 'pois', 'nbinom'
pbdist	average distance between neighbouring features, in this case in the link{biomvRhsmm} call one should only use the rank rather than the position.

Details

Be default, the hidden-semi Markov model implemented in this package uses a uniform prior for the initial sojourn distribution. However user can provide custom data from related studies to learn the prior of the sojourn distribution. The number of possible state will also be estimated from the unique level of feature type in the first meta column of xAnno if it is not a TxDb object.

Value

a list object containing the sojourn distribution parameter

type	type of the sojourn distribution
fttypes	unique levels of the types stored in the first meta column of xAnno, alphabetically sorted
J	number of possible states
\code...	distribution parameters, 'lambda' and 'shift' for 'pois'; 'size', 'mu' and 'shift' for 'nbinom'; 'scale' and 'shape' for 'gamma'

Author(s)

Yang Du

Examples

	data(encodeTP53)
	encodeTP53$gmgr # a GRanges object
	soj<-sojournAnno(encodeTP53$gmgr, soj.type='gamma')

biomvRCNS documentation built on Nov. 8, 2020, 6:49 p.m.