Usage
sciClone(vafs, copyNumberCalls=NULL, regionsToExclude=NULL, sampleNames, minimumDepth=100, clusterMethod="bmm", clusterParams="no.apply.overlapping.std.dev.condition", cnCallsAreLog2=FALSE, useSexChrs=TRUE, doClustering=TRUE, verbose=TRUE, copyNumberMargins=0.25, maximumClusters=10, annotation=NULL, doClusteringAlongMargins=TRUE, plotIntermediateResults=0)
Arguments
vafs
a list of dataframes containing variant allele fraction data for
single nucleotide variants in 5-column format:
1. chromosome 2. position 3. reference-supporting
read counts 4. variant-supporting read counts 5. variant allele
fraction (between 0-100)
copyNumberCalls
list of dataframes containing copy number segments in
4-column format: 1. chromosome 2. segment start position 3. segment
stop position 4. copy number value for that segment. Unrepresented
regions are assumed to have a copy number of 2.
regionsToExclude
Exclusion regions in 3-column format: 1. chromosome 2. window
start position 3. window stop position; Single nucleotide variants
falling into these windows will not be included in the analysis. Use
this input for LOH regions, for example.
sampleNames
vector of names describing each sample ex: ("Primary Tumor", "Relapse")
minimumDepth
threshold used for excluding low-depth variants
maximumClusters
max number of clusters to consider when choosing the component
fit to the data.
annotation
a list of positions in 3-column format 1) chromosome 2) position 3)
gene name. These will be used to annotate the cluster table, if output.
cnCallsAreLog2
boolean argument specifying whether or not the copy number
predictions are in log2 format (as opposed to being absolute copy
number designations)
useSexChrs
boolean argument to specify preference of whether (TRUE) or not
(FALSE) to use variants on sex chromosomes in the clustering
steps of the tool.
doClustering
boolean argument - if (TRUE), the tool will attempt to use clustering
to identify subclones. If (FALSE) this stage is skipped, and an
object suitable for feeding into the plotting functions is
produced.
clusterMethod
Use a different distribution for clustering. Currently available
options are 'bmm' for beta, 'gaussian.bmm' for gaussian, and
'binomial.bmm' for binomial.
clusterParams
The framework is in place to drop in different clustering methods
and provide them with additional parameters, but none of the
currently available methods take any params - this should stay NULL.
verbose
if TRUE, prints lots of output to the screen that might be useful
for debugging.
copyNumberMargins
In order to identify cleanly copy-number neutral regions, sciClone
only considers sites with a copy number of 2.0 +/- this value. For
example, if set to 0.25, regions at 2.20 will be considered cn-neutral,
and regions at, 2.30 will not.
doClusteringAlongMargins
Perform 1d clustering of each sample to facilitate certain certain
types of plotting (via sc.plot2dWithMargins())
plotIntermediateResults
output plots from intermediate steps of clustering (allows for
vizualization of cluster convergence. Generally not useful, unless
you're debugging the clustering code.