runDBSCAN-scRNAseq: runDBSCAN

In ilyessr/conclus: ScRNA-seq Workflow CONCLUS - From CONsensus CLUSters To A Meaningful CONCLUSion

runDBSCAN

Description

Run clustering iterations with selected parameters using DBSCAN.

Usage

runDBSCAN(theObject, cores=2, epsilon=c(1.3, 1.4, 1.5), minPoints=c(3, 4),
writeOutput=FALSE)

Arguments

theObject	An Object of class scRNASeq for which the count matrix was normalized and the tSNE coordinates were calculated. See ?normaliseCountMatrix and ?generateTSNECoordinates.
cores	Maximum number of jobs that CONCLUS can run in parallel. Default is 1.
epsilon	Reachability distance parameter of fpc::dbscan() function. See Ester et al. (1996) for more details. Default = c(1.3, 1.4, 1.5)
minPoints	Reachability minimum no. of points parameter of fpc::dbscan() function. See Ester et al. (1996) for more details. Default = c(3, 4)
writeOutput	If TRUE, write the results of the dbScan clustering to the output directory defined in theObject, in the sub-directory output_tables. Default = FALSE.

Details

Following the calculation of t-SNE coordinates, DBSCAN is run with a range of epsilon and MinPoints values which will yield a total of 84 clustering solutions (PCs x perplexities x MinPoints x epsilon). minPoints is the minimum cluster size which you assume to be meaningful for your experiment and epsilon is the radius around the cell where the algorithm will try to find minPoints dots. Optimal epsilon must lay one the knee of the k-NN function as shown in the "test_clustering/distance_graph.pdf".

Value

An object of class scRNASeq with its dbscanList slot updated. Also writes the clustering results in "dataDirectory/output_tables" subfolder if the parameter writeOutput is TRUE.

Author(s)

Ilyess RACHEDI, based on code by Polina PAVLOVICH and Nicolas DESCOSTES.

See Also

normaliseCountMatrix generateTSNECoordinates

Examples

## Object scr containing the results of previous steps
load(system.file("extdata/scrFull.Rdat", package="conclus"))

## Perform the clustering with dbScan
## These parameters are tweaked to fit our example data and reduce
## computing time, please consider using the default parameters or
## adjusted to your dataset.
scr <- runDBSCAN(scr, epsilon=c(380, 390, 400), minPoints=c(2,3), cores=2)

ilyessr/conclus documentation built on April 8, 2022, 1:43 p.m.