infer.clonal.models: Infer clonal structures and evolution models for multiple...
In hdng/clonevol: Clonal ordering and visualization

Description Usage Arguments Examples

Infer clonal structures and evolution models for multi cancer samples from a single patients (eg. primary tumors, metastatic tumors, xenograft tumors, multi-region samples, etc.)

infer.clonal.models(c = NULL, variants = NULL,
  cluster.col.name = "cluster", founding.cluster = NULL,
  ignore.clusters = NULL, vaf.col.names = NULL, ccf.col.names = NULL,
  vaf.in.percent = TRUE, depth.col.names = NULL, weighted = FALSE,
  sample.names = NULL, sample.groups = NULL, model = "monoclonal",
  cancer.initiation.model = NULL, subclonal.test = "bootstrap",
  cluster.center = "median", subclonal.test.model = "non-parametric",
  seeding.aware.tree.pruning = FALSE, merge.similar.samples = FALSE,
  clone.colors = NULL, random.seed = NULL, boot = NULL,
  num.boots = 1000, p.value.cutoff = NULL, sum.p.cutoff = 0.01,
  cross.p.cutoff = NULL, alpha = NULL, min.cluster.vaf = 0.01,
  score.model.by = "probability", verbose = TRUE)

`c:`	clonality analysis data frame, consisting of N+1 columns. The first column must be named 'cluster' and hold variant cluster number (ie. use number to name cluster, starting from 1,2,3... 0 is reserved for normal cell clone). The next N columns contain VAF estimated for the corresponding cluster (values range from 0 to 0.5). Either c or variants parameter is required.
`variants:`	data frame of the variants. At least cluster column and VAF or CCF columns are required. Cluster column should contain cluster identities as continuous integer values starting from 1. Either c or variants parameter is required.
`cluster.col.name:`	column that holds the cluster identity, overwriting the default 'cluster' column name
`vaf.col.names:`	names of VAF columns, either vaf.col.names or ccf.col.names is required, but not both
`ccf.col.names:`	names of CCF columns, either vaf.col.names or ccf.col.names is required, but not both
`founding.cluster:`	the cluster of variants that are found in all samples and is beleived the be the founding events. This is often the cluster with highest VAF and most number of variants
`ignore.clusters:`	clusters to ignore (not inluded in the models). Those are the clusters that are thought of as outliers, artifacts, etc. resulted from the error or bias of the sequencing and analysis. This is provided as a debugging tool
`sample.groups:`	named vector of sample groups, later clone will be colored based on the grouping of shared samples, eg. clone specific to primaries, metastasis, or shared between them. Default = NULL
`cancer.initiation.model:`	cancer evolution model to use, c('monoclonal', 'polyclonal'). monoclonal model assumes the orginal tumor (eg. primary tumor) arises from a single normal cell; polyclonal model assumes the original tumor can arise from multiple cells (ie. multiple founding clones). In the polyclonal model, the total VAF of the separate founding clones must not exceed 0.5
`subclonal.test:`	'bootstrap' = perform bootstrap subclonal test 'none' = straight comparison of already estimated VAF for each cluster provided in c
`subclonal.test.model:`	What model to use when generating the bootstrap Values are: c('non-parametric', 'normal', 'normal-truncated', 'beta', 'beta-binomial')
`min.cluster.vaf:`	the minimum cluster VAF to be considered positive (detectable cluster); default=NULL, this will be used in two places: (i): detection of positive VAF cluster, if mean/median cluster VAF is greater; if not provided (NULL), bootstrap test will be used instead to determine if cluster VAF is significantly greater/smaller than zero (using the sum.p.cutoff param) (ii): when no bootstrap model used, any cluster VAF falling below this is considered non-existed/non-detectable cluster
`cluster.center:`	median or mean
`random.seed:`	a random seed to bootstrap generation.
`merge.similar.samples:`	if a latter sample has the same tree compared with a sample processed earlier (appear first in vaf.col.names) then that sample will be removed from the tree when merging clonal evolution trees across samples. An output file *.sample-reduction.tsv will be created when plot.clonal.models is called later.
`clone.colors:`	vector of clone colors that will be used for to visualization
`seeding.aware.tree.pruning:`	only prune a sample private subclones
`score.model.by:`	model scoring scheme. Currently there are two ways to score a model (probability & metap). In probability score, models are scored and ranked by the probability that all clonal orderings result in non-negative clonal CCF. In metap model, models are scored and ranked by the combination of the max of (pvalues of individual clonal orderings) when they do not affect clonal seeding interpretation, ie. seeding clones between samples do not change drawing in the results
`weighted:`	weighted model (default = FALSE)
`depth.col.names:`	depth to be used in beta-bionmial model

data(aml1)
y = infer.clonal.models(variants = aml1$variants,
          cluster.col.name = 'cluster',
          vaf.col.names = aml1$params$vaf.col.names,
          subclonal.test = 'bootstrap',
          subclonal.test.model = 'non-parametric',
          num.boots = 1000,
          founding.cluster = 1,
          cluster.center = 'mean',
          ignore.clusters = NULL,
          min.cluster.vaf = 0.01,
          sum.p = 0.05,
          alpha = 0.05)