Ltree-class: The Ltree Class

Description

The Ltree class is the central object storing all information generated during lineage tree inference by the StemID algorithm. It comprises a number of slots for a variety of objects.

Arguments

object: An Ltree object.

Slots

sc: An SCseq object with the RaceID3 analysis of the single-cell RNA-seq data for which a lineage tree should be derived.

ldata

List object storing information on the clustering partition, the distance matrix, and the cluster centers in dimensionally-reduced input space and in two-dimensional t-sne space. Elements: lp: vector with the filtered partition into clusters after discarding clusters with cthr cells or less. pdi:matrix with the coordinates of all cells in the embedded space. Clusters with cthr transcripts or less were discarded (see function projcells). Rows are medoids and columns are coordinates. cn: data.frame with the coordinates of the cluster medoids in the embedded space. Clusters with cthr transcripts or less were discarded. Rows are medoids and columns are coordinates. m: vector with the numbers of the clusters which survived the filtering. pdil: data.frame with coordinates of cells in the two-dimensional t-SNE representation computed by RaceID3. Clusters with cthr transcripts or less were discarded. Rows are cells and columns are coordinates. cnl: data.frame with the coordinates of the cluster medoids in the two-dimensional t-SNE representation computed by RaceID3. Clusters with cthr transcripts or less were discarded. Rows are medoids and columns are coordinates.

entropy

Vector with transcriptome entropy computed for each cell.

trproj

List containing two data.frames. Elements: res: data.frame with three columns for each cell. The first column o shows the cluster of a cell, the second column l shows the cluster number for the link the cell is assigned to, and the third column h shows the projection as a fraction of the length of the inter-cluster link. Parallel projections are positive, while anti-parallel projections are negative. rma: data.frame with all projection coordinates for each cell. Rows are cells and columns are clusters. Projections are given as a fraction of the length of the inter-cluster link. Parallel projections are positive, while anti-parallel projections are negative. The column corresponding to the originating cluster of a cell shows NA.

par

List of parameters used for the StemID2 analysis.

prback

data.frame of the same structure as the trproj$res. In case randomizations are used to compute significant projections, the projections of all pdishuff randomizations are appended to this data.frame and therefore the number of rows corresponds to the number of cells multiplied by pdishuf. See function projback.

prbacka

data.frame reporting the aggregated results of the randomizations with four columns. Column n denotes the number of the randomization sample, column o and l contain the numbers of the originating and the terminal cluster, respectively, for each inter-cluster link and column count shows the number of cells assigned to this link in randomization sample n. The discrete distribution for the computation of the link p-value is given by the data contained in this object (if nmode=FALSE).

ltcoord

Matrix storing projection coordinates of all cells in the two-dimensional t-SNE space, used for visualization.

prtree

List with two elements. The first element l stores a list with the projection coordinates for each link. The name of each element identifies the link and is composed of two cluster numbers separated by a dot. The second element n is a list of the same structure and contains the cell names corresponding to the projection coordinates stored in l.

cdata

list of data.frames, each with cluster ids as rows and columns: counts data.frame indicating the number of cells on the links connecting the cluster of origin (rows) to other clusters (columns). counts.br data.frame containing the cell counts on cluster connections averaged across the randomized background samples (if nmode = FALSE) or as derived from sampling statistics (if nmode = TRUE). pv.e matrix of enrichment p-values estimated from sampling statistics (if nmode = TRUE); entries are 0 if the observed number of cells on the respective link exceeds the (1 – pethr)-quantile of the randomized background distribution and 0.5 otherwise (if nmode = FALSE). pv.d matrix of depletion p-values estimated from sampling statistics (if nmode = TRUE); entries are 0 if the observed number of cells on the respective link is lower than the pethr-quantile of the randomized background distribution and 0.5 otherwise (if nmode = FALSE). pvn.e matrix of enrichment p-values estimated from sampling statistics (if nmode = TRUE); 1- quantile, with the quantile estimated from the number of cells on a link as derived from the randomized background distribution (if nmode = FALSE). pvn.d matrix of depletion p-values estimated from sampling statistics (if nmode = TRUE); quantile estimated from the number of cells on a link as derived from the randomized background distribution (if nmode = FALSE).