RunSWNE | R Documentation |
Wrapper for running SWNE analysis
## S3 method for class 'cisTopic'
RunSWNE(
cisTopicObject,
proj.method = "sammon",
cells.use = NULL,
dist.metric = "cosine",
n.cores = 8,
hide.factors = T,
n_pull = 3,
alpha.exp = 1.25,
snn.exp = 1,
snn.k = 20,
prune.SNN = 1/15,
use.paga.pruning = T,
sample.groups = NULL,
paga.qval.cutoff = 0.001
)
## S3 method for class 'Seurat'
RunSWNE(
object,
proj.method = "sammon",
reduction.use = "pca",
cells.use = NULL,
dims.use = NULL,
genes.use = NULL,
dist.metric = "cosine",
distance.matrix = NULL,
n.cores = 8,
k,
k.range,
var.genes,
loss = "mse",
genes.embed,
hide.factors = T,
n_pull = 3,
ica.fast = T,
alpha.exp = 1.25,
snn.exp = 1,
snn.k = 10,
use.paga.pruning = T,
sample.groups = NULL,
paga.qval.cutoff = 0.001,
reduction.name = "swne",
reduction.key = "SWNE_",
return.format = "embedding",
...
)
## S3 method for class 'Pagoda2'
RunSWNE(
object,
proj.method = "sammon",
dist.metric = "cosine",
n.cores = 8,
k,
k.range,
var.genes,
loss = "mse",
genes.embed,
hide.factors = T,
n_pull = 3,
ica.fast = T,
n.var.genes = 3000,
alpha.exp = 1.25,
snn.exp = 1,
snn.k = 20,
use.paga.pruning = T,
sample.groups = NULL,
paga.qval.cutoff = 0.001
)
## S3 method for class 'dgCMatrix'
RunSWNE(
data.matrix,
proj.method = "sammon",
dist.metric = "cosine",
n.cores = 3,
k,
k.range,
var.genes,
loss = "mse",
genes.embed,
hide.factors = T,
n_pull = 3,
ica.fast = T,
alpha.exp = 1.25,
snn.exp = 1,
snn.k = 20,
use.paga.pruning = T,
sample.groups = NULL,
paga.qval.cutoff = 0.001
)
## S3 method for class 'matrix'
RunSWNE(
data.matrix,
proj.method = "sammon",
dist.metric = "cosine",
n.cores = 8,
k,
k.range,
var.genes = rownames(data.matrix),
loss = "mse",
genes.embed,
hide.factors = T,
n_pull = 3,
alpha.exp = 1.25,
snn.exp = 1
)
## S3 method for class 'dgTMatrix'
RunSWNE(
data.matrix,
proj.method = "sammon",
dist.metric = "cosine",
n.cores = 3,
k,
k.range,
var.genes = rownames(data.matrix),
loss = "mse",
genes.embed,
hide.factors = T,
n_pull = 3,
alpha.exp = 1.25,
snn.exp = 1
)
proj.method |
Method to use to project factors in 2D. Either "sammon" or "umap" |
cells.use |
Which cells to analyze (default, all cells) |
n.cores |
Number of cores to use (passed to FindNumFactors) |
hide.factors |
Hide factors when plotting SWNE embedding |
n_pull |
Maximum number of factors "pulling" on each sample |
alpha.exp |
Increasing alpha.exp increases how much the NMF factors "pull" the samples (passed to EmbedSWNE) |
snn.exp |
Decreasing snn.exp increases the effect of the similarity matrix on the embedding (passed to EmbedSWNE) |
snn.k |
Changes the number of nearest neighbors used to build SNN (passed to CalcSNN) |
prune.SNN |
The minimum fraction of shared nearest neighbors (smaller values are set to zero) |
use.paga.pruning |
Use PAGA graphs to prune |
sample.groups |
Clusters to use for PAGA (default is to do a de-novo clustering) |
paga.qval.cutoff |
q-value cutoff for significant shared edges between clusters |
object |
A Seurat, Pagoda2 or cisTopicObject object with normalized data |
reduction.use |
Which dimensional reduction (e.g. PCA, ICA) to use for the tSNE. Default is PCA. |
dims.use |
Which dimensions to use as input features |
genes.use |
If set, run the SWNE on this subset of genes (instead of running on a set of reduced dimensions). Not set (NULL) by default |
distance.matrix |
If set, runs tSNE on the given distance matrix instead of data matrix (experimental) |
k |
Number of NMF factors (passed to RunNMF). If none given, will be derived from FindNumFactors. |
k.range |
Range of factors for FindNumFactors to iterate over if k is not given |
var.genes |
vector to specify variable genes. Will infer from Seurat or use full dataset if not given. |
loss |
loss function to use (passed to RunNMF) |
genes.embed |
Genes to add to the SWNE embedding |
ica.fast |
Whether to run SVD before ICA initialization |
reduction.name |
dimensional reduction name, specifies the position in the object$dr list. swne by default |
reduction.key |
dimensional reduction key, specifies the string before the number for the dimension names. SWNE_ by default |
return.format |
format to return ("seurat" object or raw "embedding") |
n.var.genes |
Number of variable genes to use |
data.matrix |
a data matrix (genes x cells) which has been pre-normalized |
batch |
Vector of batch IDs to regress away |
dist.use |
Similarity function to use for calculating factor positions (passed to EmbedSWNE). Options include pearson (correlation), IC (mutual information), cosine, euclidean. |
A list of factor (H.coords) and sample coordinates (sample.coords) in 2D
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.