The proportion_scaled_cnv of all cells in observation group were identical??

[fengwei-li]

Hi, I did the inforcnv calculation, however, the results I got indicated that The proportion_scaled_cnv of all cells(in observation group) were identical. Moreover, for a particular chromosome, only one score(or 0) was given, which was not reasonable(cnv.txt). I'm very very confused.

infercnv_obj <- CreateInfercnvObject(raw_counts_matrix = count,
                                     annotations_file = '6.cell_group.txt',
                                     delim="\t",
                                     gene_order_file = '6.gene_order.txt',
                                     ref_group_names = "Fibroblasts")
n <- 'inferCNV_20211121'
infercnv_obj = infercnv::run(infercnv_obj,
                             cutoff=0.1, 
                             out_dir=n, 
                             cluster_by_groups=F,
                             hclust_method="ward.D2",
                             HMM=T,
                             denoise = T,
                             num_threads = 4)
chol_CAF_cnv = infercnv::add_to_seurat(
  infercnv_output_path="G:/Analysis_data/Step6/inferCNV_20211121/", seurat_obj=chol_CAF,top_n=10)

cnv.txt

[GeorgescuC]

Hi @fengwei-li ,

The results from add_to_seurat are based on the HMM predictions. Since you did not use the subclustering option, there is a single cluster containing all your observation cells, and the whole cluster receives a single set of predictions, as you can check if you look at the HMM figures. Seeing your cells are clearly from multiple clonal populations, You should use the analysis_mode="subclusters" option to get meaningful HMM predictions and metric export from add_to_seurat.

Regards, Christophe.