ti_raceid_stemid: RaceID / StemID

In dynverse/dynmethods: A collection of trajectory inference methods

RaceID / StemID

Description

Will generate a trajectory using RaceID / StemID.

This method was wrapped inside a container. The original code of this method is available here.

Usage

ti_raceid_stemid(
  knn = 10L,
  ccor = 0.4,
  metric = "pearson",
  sat = TRUE,
  samp = 1000L,
  cln = 30L,
  clustnr = 30L,
  bootnr = 50L,
  FUNcluster = "kmedoids",
  probthr = 0.001,
  outminc = 5L,
  outlg = 2L,
  outdistquant = 0.95,
  initial_cmd = TRUE,
  perplexity = 30L,
  cthr = 5L,
  nmode = TRUE,
  projcells_knn = 3L,
  fr = FALSE,
  pdishuf = 500L,
  fast = FALSE,
  pthr = 0.01,
  scthr = 0.2
)

Arguments

knn	Number of nearest neighbors used to infer corresponding cell types in different batches. Domain: U(5, 50). Default: 10. Format: integer.
ccor	Correlation coefficient used as a treshhold for determining genes correlated to eachother. Domain: U(0, 1). Default: 0.4. Format: numeric.
metric	Distances are computed from the filtered expression matrix after optional feature selection, dimensional reduction, and/or transformation (batch correction). Domain: pearson, spearman, logpearson, euclidean. Default: pearson. Format: character.
sat	If TRUE, then the number of clusters is determined based on finding the saturation point of the mean within-cluster dispersion as a function of the cluster number. If FALSE, then cluster number needs to be given as cln. Default: TRUE. Format: logical.
samp	Number of bootstrapping runs for clusterboot. Domain: e^U(3.91, 9.21). Default: 1000. Format: integer.
cln	Number of clusters to be used. If sat is TRUE, this number is inferred by the saturation criterion. Domain: U(10, 100). Default: 30. Format: integer.
clustnr	Maximum number of clusters for the derivation of the cluster number by the saturation of mean within-cluster-dispersion. Domain: U(10, 100). Default: 30. Format: integer.
bootnr	Number of booststrapping runs for clusterboot. Domain: U(20, 100). Default: 50. Format: integer.
FUNcluster	Clustering method used by RaceID3. Domain: kmedoids, kmeans, hclust. Default: kmedoids. Format: character.
probthr	Outlier probability threshold for a minimum of outlg genes to be an outlier cell. This probability is computed from a negative binomial background model of expression in a cluster. Domain: e^U(-11.51, 0.00). Default: 0.001. Format: numeric.
outminc	Minimal transcript count of a gene in a clusters to be tested for being an outlier gene. Domain: U(0, 100). Default: 5. Format: integer.
outlg	Minimum number of outlier genes required for being an outlier cell. Domain: U(0, 100). Default: 2. Format: integer.
outdistquant	Real number between zero and one. Outlier cells are merged to outlier clusters if their distance smaller than the outdistquant-quantile of the distance distribution of pairs of cells in the orginal clusters after outlier removal. Domain: U(0, 1). Default: 0.95. Format: numeric.
initial_cmd	If TRUE, then the t-SNE map computation is initialized with a configuration obtained by classical multidimensional scaling. Default: TRUE. Format: logical.
perplexity	Perplexity of the t-SNE map. Domain: U(5, 100). Default: 30. Format: integer.
cthr	Clusters to be included into the StemID2 analysis must contain more than cthr cells. D. Domain: U(1, 25). Default: 5. Format: integer.
nmode	If TRUE, then a cell of given cluster is assigned to the link to the cluster with the smallest average distance of the knn nearest neighbours within this cluster. Default: TRUE. Format: logical.
projcells_knn	See nmode. Domain: U(3, 20). Default: 3. Format: integer.
fr	Use Fruchterman-Rheingold layout instead of t-SNE for dimensional-reduction representation of the lineage tree. Default: FALSE. Format: logical.
pdishuf	Number of randomizations of cell positions for which to compute projections of cells on inter-cluster links. Domain: e^U(2.30, 9.21). Default: 500. Format: integer.
fast	If TRUE and nmode is FALSE cells will still be assigned to links based on maximum projections but a fast approximate background model will be used to infer significance. The function will do nothing in this case. Default: FALSE. Format: logical.
pthr	P-value cutoff for link significance. This threshold is applied for the calculation of link scores reflecting how uniformly a link is occupied by cells. Domain: e^U(-11.51, 0.00). Default: 0.01. Format: numeric.
scthr	Score threshold for links to be shown in the graph. Domain: U(0, 1). Default: 0.2. Format: numeric.

Value

A TI method wrapper to be used together with infer_trajectory

References

Grün, D., Muraro, M.J., Boisset, J.-C., Wiebrands, K., Lyubimova, A., Dharmadhikari, G., van den Born, M., van Es, J., Jansen, E., Clevers, H., de Koning, E.J.P., van Oudenaarden, A., 2016. De Novo Prediction of Stem Cell Identity using Single-Cell Transcriptome Data. Cell Stem Cell 19, 266–277.

dynverse/dynmethods documentation built on Jan. 18, 2024, 4:44 a.m.