Algorithm settings used in Lai et al. (2005)

CGHseg (default settings -- note that the sign of S is different between
        the value given in Picard et al. and the value given in the MATLAB
	code... the implementations are actually identical because a negative
	sign was implicitly removed in the MATLAB code):
Km = 20, S = 0.5


Quantreg (following Eilers and de Menezes's tuning recommendation):
lambda = 1.5, tau = 0.5


CLAC (default settings):
windowsize = 5, targetFDR = 0.01


GLAD (default settings):
mediancenter = FALSE, smoothfunc = "lawsglad", bandwidth = 10, round = 2,
model = "Gaussian", lkern = "Exponential", qlambda = 0.999, base = FALSE,
sigma = NULL, lambdabreak = 8, lambdacluster = 8, lambdaclusterGen = 40,
type = "tricubic", param = c(d = 6), alpha = 0.001, msize = 5,
method = "centroid", nmax = 8


CBS (default settings):
alpha = 0.01, nperm = 10000, window.size = NULL, overlap = 0.25, trim = 0.025,
undo.splits = "none", undo.prune = 0.05, undo.SD = 3


HMM (default settings):
vr = 0.01, maxiter = 100, aic = TRUE, bic = TRUE, delta = NA, model.use = 1,
minDiff = 0.25, maxmadUse = 0.3, maxmedUse = 0.5, maxState = 3, minClone = 20,
factor = 5, maxClones = 1, maxLen = 1000, absValSingle = 1, absValRegion = 1,
diffVal1 = 1, diffVal2 = 0.5, maxSize = 10000, pChrom.min = 0.9,
medChrom.min = 0.1)


Wavelet (following results of Hsu et al.):
n.levels = 3, hard = FALSE


Lowess (our implementation):
w = 10 (meaning use a window size of 10), iter = 3, delta = 0


ChARM (default settings):
meanPcutoff = 0.01, signPcutoff = 0.001, signTest = 1


GA (turn off amplification cutoff threshold):
amplificationThreshold = 100 (effectively turns off the threshold parameter
                              because the default checkbox code has a bug)


ACE (our selection since an target FDR was not given):
targetFDR = 0.001