Single result vs. merged result

[DLEE-BARC]

Hi, I am using cnvpytor to analyze over 90 samples.

I removed PCR duplicates using gatk.
I called CNVs for each sample by following the steps:

1. -rd; 2) -hist; 3) -partition; 4) -snp (import snp data); 5) -pileup; 6) -mask_snps; 7) -baf; 8) -call (.tsv file).

Once I was done with these steps for each sample, I merged and filtered the .pytor files in -view mode as recommended here (/~https://github.com/abyzovlab/CNVpytor/blob/master/examples/merging.md)

I finally printed the 'merged_calls' in an Excel format.

However, as you can see, the results between single results and merged results showed different outcomes.
How can I interpret this result?

This is sample 1 result.

I am looking for your help! Thank you so much!

[arpanda]

Hi @DLEE-BARC,
In the Excel file, the CNV values are rounded to the nearest integer copy number for each sample. However, if you click on a cell, you can view the exact CNV value. For example, in Sample 1, the first result has an actual value of 0.1369, which is rounded to 0.

If you could share a few rows from Sample 2, it would be easier to explain.

-Arijit

[DLEE-BARC]

Thank you for the reply, @arpanda

I am curious about the CNV values, for example, in Chr01:14001-109000 in the merged file (excel file).
It doesn't appear in the single file of sample 1, but there is a CNV value of ~3, which was considered duplication in the merged file.

How can I interpret that kind of inconsistency after merging multiple files?

Thank you so much!

[arpanda]

To call CNVs, a region must meet certain criteria. In Sample 1, the evidence for the region Chr01:14001-109000 was insufficient, whereas in Sample 2, it was sufficient. This difference may result from factors like low sequencing coverage or noise. During merging, a CNV call from one sample is used, and genotyping is performed only for that called region in other samples to determine CNV values.