segmentByGFLars: Group fused Lars segmentation (low-level)
In jointseg: Joint segmentation of multivariate (copy number) signals

Description Usage Arguments Details Value Note Author(s) References See Also Examples

Low-level function for multivariate fused Lars segmentation (GFLars)

1 2	segmentByGFLars(Y, K, weights = defaultWeights(nrow(Y)), epsilon = 1e-09, verbose = FALSE)

`Y`	A `n*p` matrix of signals to be segmented
`K`	The number of change points to find
`weights`	A `(n-1)*1` vector of weights for the weigthed group fused Lasso penalty. See Details.
`epsilon`	Values smaller than epsilon are considered null. Defaults to `1e-9`.
`verbose`	A `logical` value: should extra information be output ? Defaults to `FALSE`.

This function recrusively looks for the best candidate change point according to group-fused LARS. This is a low-level function. It is generally advised to use the wrapper doGFLars which also works on data frames, has a convenient argument stat, and includes a basic workaround for handling missing values.

See also jointSeg for combining group fused LARS segmentation with pruning by dynamic programming (pruneByDP).

See PSSeg for segmenting genomic signals from SNP arrays.

The default weights √{n/(i*(n-i))} are calibrated as suggested by Bleakley and Vert (2011). Using this calibration, the first breakpoint maximizes the likelihood ratio test (LRT) statistic.

A list with elements:

`bkp`	A vector of `k` candidate change-point positions
`lambda`	The estimated lambda values for each change-point
`mean`	A vector of length `p`, the mean signal per column
`value`	A `i x p` matrix of change-point values for the first i change-points
`c`	\hat{c}, a `n-1 x K` matrix

This implementation is derived from the MATLAB code by Vert and Bleakley: http://cbio.ensmp.fr/GFLseg.

Morgane Pierre-Jean and Pierre Neuvial

Bleakley, K., & Vert, J. P. (2011). The group fused lasso for multiple change-point detection. arXiv preprint arXiv:1106.4199.

Vert, J. P., & Bleakley, K. (2010). Fast detection of multiple change-points shared by many signals using group LARS. Advances in Neural Information Processing Systems, 23, 2343-2351.

PSSeg, jointSeg, doGFLars, pruneByDP

p <- 2
trueK <- 10
sim <- randomProfile(1e4, trueK, 1, p)
Y <- sim$profile
K <- 2*trueK
res <- segmentByGFLars(Y, K)
print(res$bkp)
print(sim$bkp)
plotSeg(Y, res$bkp)