View source: R/projectedNeighbors.R
projectedNeighbors | R Documentation |
Computes composite distances between all cell pairs and returns k-nearest neighbors and edge weights needed to build VeloViz graph.
projectedNeighbors( observed, projected, k, distance_metric = "L2", similarity_metric = "cosine", distance_weight = 1, distance_threshold = 1, similarity_threshold = -1 )
observed |
PCs (rows) x cells (columns) matrix of observed transcriptional state projected into PC space |
projected |
PCs (rows) x cells (columns) matrix of projected transcriptional states. Cells should be in same order as in |
k |
Number of nearest neighbors to assign each cell |
distance_metric |
Method to compute distance component of composite distance. "L1" or "L2", default = "L2" |
similarity_metric |
Method to compute similarity between velocity and cell transition matrices. "cosine" or "pearson", default = "cosine" |
distance_weight |
Weight of distance component of composite distance, default = 1 |
distance_threshold |
quantile threshold for distance component above which to remove edges, default = 1 i.e. no edges removed |
similarity_threshold |
similarity threshold below which to remove edges, default = -1 i.e. no edges removed |
kNNs
cells (rows) x k (columns) matrix of indices of each cell's nearest neighbors computed based on composite distance. Edges removed based on distance or similarity threshold will be NA.
edge_weights
cells (rows) x k (columns) matrix of edge weights computed based on composite distance. Edges removed based on distance or similarity threshold will be NA.
all_dists
cells x cells matrix of all pairwise composite distances
dist_comp
components of composite distance: invDist
distance component, negSim
similarity component
graphViz
data(vel) curr <- vel$current proj <- vel$projected projectedNeighbors(curr, proj, 10)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.