Description Usage Arguments Value References Examples
This method clusters sample data using a two-step approach. In the first step it converts the high dimensional data into a knn-nearest neighbor graph, thus connecting samples that have a similar profile. The edges in knn-nearest neighbors graph are given weights according to the Jaccard similarity coefficient of the neighbors of each node. In the second step the Louvaine algorithm, which was developed to find communities in social networks, is applied to this graph in order to find samples that share similar spaces in multi-dimensional space.
1 2 |
X |
an M-by-N numeric matrix (or any object that can
be coerced to a matrix) holding N samples in columns,
where each column has M values. X can also be an object
of class |
method |
is a string that is passed to the
|
D |
is a dist object. If provided then |
knn |
a numeric value indicating the number of nearest neighbors to use in the initial neighborhood (default=NULL uses nclass.Sturges) |
G |
an N-by-N adjacency matrix on which modularity will be maximized.
If provided |
C |
an optional vector of length N with the initial numeric cluster assignments (default is for each sample to belong to its own singleton cluster) |
repeats |
a numeric value for how many times to run the Louvain algorithm in order to find a maximal local maximum (default=50). |
verbose |
a boolean vaue indicating whether to print progress messages (default=TRUE) |
a list holding three values: C
- a numeric vector
of length N indicating the cluster number. G
- an
N-by-N matrix representing the Hadamrd graph. Q
-
the modularity provided by G
and C
.
Levine et al, "Data-Driven Phenotypic Dissection of AML Reveals Progenitor-like Cells that Correlate with Prognosis", Cell, Volume 162, Issue 1, 2 July 2015, Pages 184-197
1 2 3 |
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.