Monocle plot_pseudotime_heatmap : Error in cutree(tree, cutree_n) : elements of 'k' must be between 1 and 3
0
0
Entering edit mode
A.H • 0
@ah-13492
Last seen 4.2 years ago

I immediately admit that I am an expert in R, or scripting in general for that matter, so I hope this is an easy question: 
I do not know how to include a decent example dataset for this so I try to explain the problem

I am analyzing single-cell RNA-seq data, using the Monocle (Bioconductor) package. This package allows you to order cells in pseudotime based on their transcriptome,  thereby providing insight into transcriptome dynamics during development in pseudotime. I am trying to use the command plot_pseudotime_heatmap to generate a heatmap for a fitted linear model on the gene-expression pattern of a number of disease-markers in pseudotime.

Though for some genesets I get the following error when plotting the heatmap: 
Error in cutree(tree, cutree_n) : elements of 'k' must be between 1 and 3

I cannot figure out what k is (I thought the number of cluster to be formed, but why does this need to be between 1 and 3...), or which setting to change to overcome this error:


Default settings for plot_pseudotime_heatmap():

plot_pseudotime_heatmap(cds_subset, cluster_rows = TRUE,
  hclust_method = "ward.D2", num_clusters = 6, hmcols = NULL,
  add_annotation_row = NULL, add_annotation_col = NULL,
  show_rownames = FALSE, use_gene_short_name = TRUE,
  norm_method = c("vstExprs", "log"), scale_max = 3, scale_min = -3,
  trend_formula = "~sm.ns(Pseudotime, df=3)", return_heatmap = FALSE,
  cores = 1)

As far as I understand, in this package, the command plot_pseudotime_heatmap, first orders the cells according to their pseudotime in 'development', next fits a linear model to the expression pattern per gene (with lm.fit()), and finally transforming this data into a heatmap (using a variant of pheatmap() ), where the pseudotime is split over 100 pseudotime-bins.


I have the difficulty of working with a very rare cellpopulation, causing some for the disease-marker genes to be expressed in only 1 or a few cells and with only a few reads per gene (cut-off for expressed-genes is  >=1read in >=1 cell). I understand that trying to fit a linear model on these genes will be impossible, and that this gene will get excluded from the heatmap and I will get the error: <simpleError in lm.fit(X.vlm, y = z.vlm, ...): NA/NaN/Inf in 'y'>. But why can I sometimes continue with the remaining genes, but sometimes still get the above error on the remaining genes??

Many thanks for any help I can get!

 

R monocle • 922 views
ADD COMMENT

Login before adding your answer.

Traffic: 198 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6