lb15
commented
4 years ago I tried to run find_gene_modules() with a cds and gene list created before I upgraded R/Monocle3, and it worked. So it may be something with how i have created this cds object. One difference is that the cds_unbias contains cells across partitions that I want to analyze. I upgraded to the development monocle3 because of the fix in #275 . In the older version of monocle3, I was manually replacing the partition identifiers with 1 for all cells, to create one partition.
Here is how I built my cds:
exprs<- GetAssayData(seur_ob, slot="counts",assay = "RNA")
## phenodata
pheno.data = seur_ob@meta.data
## feature data
genes <- data.frame(gene_short_name = rownames(exprs))
rownames(genes) <- rownames(exprs)
cds <- new_cell_data_set(exprs,
cell_metadata = pheno.data,
gene_metadata = genes)
################### PROCESS DATA #########################
set.seed(123)
cds=preprocess_cds(cds, num_dim=20)
cds=align_cds(cds, preprocess_method="PCA",alignment_group="batch")
cds <- reduce_dimension(cds, umap.fast_sgd = FALSE,preprocess_method = "Aligned")
cds <- cluster_cells(cds,cluster_method="leiden",random_seed=123)
cds <- learn_graph(cds,use_partition = F ,learn_graph_control = list(minimal_branch_len=25))
cds=order_cells(cds)
cds_sub = cds[,colData(cds)$Merged_MCC_clusters == "MCC"]
##code to produce vector of genes expressed in at least 10% of cells in any cluster
cds_unbias=cds_sub[pct.above10$genes,]
### Regression analysis
gene_fits <- fit_models(cds_unbias, model_formula_str = "~pseudotime")
fit_coefs <- coefficient_table(gene_fits)
pseudotime_terms <- fit_coefs %>% filter(term == "pseudotime")
pseudotime_sig <- pseudotime_terms %>% filter (q_value < 0.05) %>%
select(gene_short_name, term, q_value, estimate)
sig_genes = pseudotime_sig$gene_short_name
modules=find_gene_modules(cds_unbias[as.character(sig_genes),])
afaissa
brgew
If this is a question and not a bug report or enhancement request, please post to our google group at https://groups.google.com/forum/#!forum/monocle-3-users
Describe the bug I'm running find_gene_modules to group genes changing over pseudotime into modules. I get an error in RANN::nn2(). I've tried with a full cds and a subsetted cds (subsetted on both cells and genes) but both produce the error. I've also tried specifying monocle3::find_gene_modules() but that is also producing the same error.
To Reproduce
traceback() After the error, run traceback() in R and post the output:
Expected behavior A clear and concise description of what you expected to happen.
Screenshots If applicable, add screenshots to help explain your problem.
sessionInfo():
Additional context I'm using the development branch of monocle3. I recently upgraded to R 4.0.0 and the development version of monocle3. Prior to this I did not get the error in find_gene_modules. Thanks!!