numbat icon indicating copy to clipboard operation
numbat copied to clipboard

how to remove cnv like noise

Open hanjun98 opened this issue 1 year ago • 2 comments

Hi, I have a quetion about removing cnv like noise

First, I use Run_numbat(function) with option like this

out = run_numbat( count_mat, # gene x cell integer UMI count matrix ref_hca, # reference expression profile, a gene x cell type normalized expression level matrix df_allele, # allele dataframe generated by pileup_and_phase script genome = "hg38", t = 1e-5, ncores = 5, plot = TRUE, out_dir = paste0('/mnt/gmi-l1/_90.User_Data/dhthxkr/MDS/',sample,'/Numbat') )

And then, I got this result(panel_2.png) panel_2

but, I think CNV in chr12 is noise, so I wanna remove them, and then use Run_numbat(function) with segs_loh Because Your documentation explain that segs_loh: dataframe Segments of clonal LOH to be excluded, so I think this option make them to remove noise.

out = run_numbat( count_mat, # gene x cell integer UMI count matrix ref_hca, # reference expression profile, a gene x cell type normalized expression level matrix df_allele, # allele dataframe generated by pileup_and_phase script genome = "hg38", t = 1e-5, ncores = 10, plot = TRUE, segs_loh = segs_loh, out_dir = paste0('/mnt/gmi-l1/_90.User_Data/dhthxkr/MDS/',sample,'/Numbat_v2') )

스크린샷 2024-01-29 오후 1 32 47

But, I got this result(panel_2.png). panel_2

It have little bit changes, but still have CNV in chr12. so I'm curious how to remove CNV like noise and why couldn't remove CNV in chr12, when I use segs_loh(option)

Thank you for reading

hanjun98 avatar Jan 29 '24 04:01 hanjun98

Hello, segs_loh option is meant for very pure tumors (almost no normal cells) that have clonal deletions in certain regions, which can be excluded using segs_loh in order to more accurately find the diploid baseline.

If you want to remove CNVs that have weak support you can adjust min_LLR or max_entropy. See https://kharchenkolab.github.io/numbat/reference/run_numbat.html

teng-gao avatar Jan 29 '24 20:01 teng-gao

Thank you for reply! I have a one question? In my data, I think 12(del) is noise, but If I adjust min_LLR or max_entropy to remove 12(del), 7(del,True positive in WES) is also removed. Because 7(del)'s avg_entrophy is higher than 12(del)'s and 7(del)'s LLR is lower than 12(del)'s. So can I make remove 12(del) only in this case?

hanjun98 avatar Feb 05 '24 05:02 hanjun98