findStretches {CAGEfightR} | R Documentation |
Finds stretches or groups of clusters along the genome, where each cluster is within a certain distance of the next. Once stretches have been identified, the average pairwise correlation between all clusters in the stretch is calculated. A typical use case is to look for stretches of enhancers, often refered to as "super enhancers".
findStretches(object, ...) ## S4 method for signature 'GRanges' findStretches(object, mergeDist = 10000L, minSize = 3L) ## S4 method for signature 'RangedSummarizedExperiment' findStretches( object, inputAssay, mergeDist = 10000L, minSize = 3L, corFun = cor, ... )
object |
GRanges or RangedSummarizedExperiment: Clusters, possibly with expression for calculating correlations. |
... |
additional arguments passed to methods or ultimately corFun. |
mergeDist |
integer: Maximum distance between clusters to be merged into stretches. |
minSize |
integer: Minimum number of clusters in stretches. |
inputAssay |
character: Name of assay holding expression values (if object is a RangedSummarizedExperiment) |
corFun |
function: Function for calculating correlations. Should behave and produce output similar to cor(). |
A GRanges containing stretches with number of clusters and average pairwise correlations calculated. The revmap can be used to retrieve the original clusters (see example below.)
Other Spatial functions:
findLinks()
,
trackLinks()
# Calculate TPM values for bidirectional clusters data(exampleBidirectional) BCs <- calcTPM(exampleBidirectional) # Find stretches pearson_stretches <- findStretches(BCs, inputAssay="TPM") # Use Kendall instead of pearson and require bigger stretches kendall_stretches <- findStretches(BCs, inputAssay="TPM", minSize=5, method="kendall") # Use the revmap to get stretches as a GRangesList grl <- extractList(rowRanges(BCs), kendall_stretches$revmap) names(grl) <- names(kendall_stretches)