mergeStates: Function to merge states based on their state means

In snapCGH: Segmentation, normalisation and processing of aCGH data

Description

mergeStates takes the output of a segmentation algorithm in the form of a SegList and iteratively merges the states with means closer than a supplied threshold.

Usage

mergeStates(segList, MergeType = 1, pv.thres=0.0001, ansari.sign=0.01, minDiff = 0.25)

Arguments

segList	Object of class SegList.
MergeType	Select either 1 or 2. 1 uses a new merging algorithm developed by Hanni Willenbrock and Jane Fridlyand.
pv.thres	Significance threshold for Wilcoxon test for level merging. Used when MergeType = 1.
ansari.sign	Significance threshold for Ansari-Bradley test. Used when MergeType = 1.
minDiff	The states whose predicted values are less than minDiff apart are merged into one state and all the predicted values are recomputed. Used when MergeType = 2.

Details

This function is intended to reduce effect of the possible small magnitude technological artifacts on the structure determination.

Value

A SegList object is returned with the merged states stored in the pred list.

Author(s)

Jane Fridlyand

References

Application of Hidden Markov Models to the analysis of the array CGH data, Fridlyand et.al., JMVA, 2004

See Also

SegList, runHomHMM, runGLAD, runDNAcopy

snapCGH documentation built on Nov. 8, 2020, 5:31 p.m.