Issue with extract_gene_trajectory

[yc508]

I have trouble importing cal_ot_mat_from_numpy using reticulate so I have converted my Seurat object to anndata with Seurat Disk and use Gene Trajectory in python. And I have stuck at the step extract_gene_trajectory. The error message I receive is as following. Could you help me out with this issue? Or could you should me where can I find the gene_distance_cal.py so I can continute using Gene Trajectory in R. Thank you very much!

ValueError: Cannot apply_along_axis when any iteration dimensions are 0

[RihaoQu]

Hi,

This error message is very vague. It seems the input matrices might not be in the right shape. Can you check the dimensions?

The easiest way is probably to use R and bypass reticulate. You can export the matrices to a folder and run the Python code from command-line. In this way there is no need to convert to Anndata.

From R

dir.path <- 'temporary_path'
dir.create(dir.path, recursive=T)
write.table(cg_output[["features"]], paste0(dir.path, "gene_names.csv"), row.names = F, col.names = F, sep = ",")
write.table(cg_output[["graph.dist"]], paste0(dir.path, "ot_cost.csv"), row.names = F, col.names = F, sep = ",")
Matrix::writeMM(Matrix(cg_output[["gene.expression"]], sparse = T), paste0(dir.path, "gene_expression.mtx"))

Run from command line (assuming gene trajectory is installed in the Python environment)

python  -m gene_trajectory.compute_gene_distance_cmd --path 'temporary_path'

Load back results

gene.dist.mat <- LoadGeneDistMat(dir.path, file_name = "emd.csv")

[yc508]

Hi,

Thank you very much for your help! It works now.