cluster assembled transcripts and plot the results

Description

This is an experimental, first-pass function that clusters assembled transcripts based on their overlap percentage, then plots and colors the transcript clusters.

Usage

1
2
3
plotLatentTranscripts(gene, gown, method = c("hclust", "kmeans"), k = NULL,
  choosek = c("var90", "thumb"), returncluster = TRUE,
  labelTranscripts = TRUE, ...)

Arguments

gene

string, name of gene whose transcripts should be clustered (e.g., "XLOC_000001")

gown

object of class ballgown being used for analysis

method

clustering method to use. Currently can choose from hierarchical clustering (hclust) or K-means (kmeans). More methods are in development.

k

number of transcripts clusters to use. By default, k is NULL and thus is chosen using a rule of thumb, but providing k overrides those rules of thumb.

choosek

if k is not provided, how should the number of clusters be chosen? Must be one of "var90" (choose a k that explains 90 percent of the observed variation) or "thumb" (k is set to be approximately sqrt(n), where n is the total number of transcripts for gene)

returncluster

if TRUE (as it is by default), return the results of the call to clusterTrancsripts so the data is available for later use. Nothing is returned if FALSE.

labelTranscripts

if TRUE (as it is by default), print transcript IDs on the y-axis

...

other arguments to pass to plotTranscripts

Value

if returncluster is TRUE, the transcript clusters are returned as described in clusterTranscripts. A plot of the transcript clusters is also produced, in the style of plotTranscripts.

Author(s)

Alyssa Frazee

See Also

clusterTranscripts, plotTranscripts

Examples

1
2
data(bg)
plotLatentTranscripts('XLOC_000454', bg, method='kmeans', k=2)

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.