CrossICC: CrossICC: Automatically Aggregating and Summarizing...

In bioinformatist/CrossICC: An Interactive Consensus Clustering Framework for Multi-platform Data Analysis

Description

CrossICC: Automatically Aggregating and Summarizing Bioinformatics Results for Interactive Report.

The Main Function of the package

Usage

CrossICC(..., study.names, filter.cutoff = 0.5, fdr.cutoff = 0.001,
  output.dir = "~", max.K = 10, max.iter = 20, rep.runs = 1000,
  n.platform = 2, pItem = 0.8, pFeature = 1, clusterAlg = "hc",
  distance = "euclidean", sil.filter = "soft",
  heatmap.order = "up.based", com.mode = "overlap", cc.seed = NULL,
  cluster.cutoff = 0.05, ebayes.cutoff = 0.1, ebayes.mode = "up",
  cross = "cluster", supercluster.method = "hclust",
  skip.merge.dup = TRUE, skip.mm = FALSE, skip.mfs = FALSE,
  use.shiny = FALSE, overwrite = FALSE)

Arguments

...	all datasets (matrices is better)
study.names	a vector containing all study names
filter.cutoff	low variability (median absolute deviation (MAD)) cutoff threshold, default is 0.5.
fdr.cutoff	cutoff value during fdr filtering.
output.dir	the results' output directory.
max.K	the maximum cluster number of ConsensusClusterPlus. Default is 10, but was set as number of samples when there're less than 10 samples.
max.iter	the maximum number of iterations.
rep.runs	number of subsamples during clustering.
n.platform	to filter the signature with it's meta-cluster group in platforms. That is, if the parameter is set to 2 (default), the signature (like hgnc symbol ESR1) in a certain meta-cluser (like K1) must exists more than 2 times among data of all platforms; otherwise, it will not be reported.
pItem	proportion of items to sample during clustering.
pFeature	proportion of features to sample during clustering.
clusterAlg	cluster algorithm. Could be 'hc' heirarchical (hclust), 'pam' for paritioning around medoids, 'km' for k-means upon data matrix, 'kmdist' for k-means upon distance matrices (former km option), or a function that returns a clustering.
distance	Could be 'pearson': (1 - Pearson correlation), 'spearman' (1 - Spearman correlation), 'euclidean', 'canberra', 'minkowski" or custom distance function.
sil.filter	silhouetee width filtering mode. Could be "soft" or "hard". If "hard", all negtive silhouetee width value will be set to 0. Default is "soft" (to do nothing).
heatmap.order	gene order for heatmaps. Default is "up.based", with which genes will be arranged as up-regulated order in meta-clusters across all matrices. Or can be set to "concordant" for all in same order.
com.mode	mode for choose common features when pre-processing data. Could be "overlap" (use intersection, default) or "merge" (keep all features).
cc.seed	sets random seed for reproducible results.
cluster.cutoff	cutoff value during determining cluster numbers.
ebayes.cutoff	p-value cutoff when select differentially expressed probes.
ebayes.mode	'up' or 'both'. Choose only up-regulated genes or all differentially expressed genes when determining MDEGs. default is 'up'
cross	object type when determining meta-cluster. Could be "cluster" for clusters by ConsencusClusterPlus, "sample" for samples or "none" (only used for single dataset).
supercluster.method	method for super-clustering. Default is 'hclust', can also be 'kmeans'.
skip.merge.dup	skip merge multiple probes for one gene (duplicates) or not. Default is TRUE (it is highly recommended that user has their data pre-processed well).
skip.mm	skip MergeMaid processing or not. Default is FALSE (not skip).
skip.mfs	by default, the datasets will be normalized at the start, and the genes or features that have no or few contributions to the final clusters will be filtered out. To skip this process, you can set this parameter to TRUE. Only try when you're sure that you're working with pre-processed datasets.
use.shiny	if TRUE, a shiny app will appear after running this main function. Note: You must keep output.dir with default value '~' for using shiny app.
overwrite	if user allow overwrite result file? Default is FALSE.

Value

A nested list with iteration time as its name and list containing consensus cluster, gene signature and balanced cluster as its value.

See Also

ConsensusClusterPlus

Examples

data("demo.platforms")
CrossICC.obj <- CrossICC(demo.platforms, skip.mfs = TRUE, max.iter = 1, overwrite = TRUE, output.dir = tempdir())

bioinformatist/CrossICC documentation built on Feb. 3, 2022, 8:58 a.m.