problem about hypothesis test in cnv analysis based TCGA about geotcgadata

[huerqiang]

1. differential_cnv was written a long time ago, and diff_CNV is an improved version. I recommend diff_CNV as the standard process in the documentation of the package. The results of both functions are reasonable.
2. The reason why differential_cnv uses 0.5 as the threshold is based on such grouping:
   Less than 0.5: homozygous loss.
   -0.5 to 1.5: hemizygotic deletion.
   1.5 to 2.5: normal copy number .
   2.5 to 3.5: Copy number amplification
   More than 3.5: high-level amplification.
3. diff_CNV directly processes the Gene Level Copy Number Scores files obtained by TCGA, and its grouping method is as follows https://docs.gdc.cancer.gov/Data/Bioinformatics_Pipelines/CNV_Pipeline/.
4. The fisher exact test is more accurate than the chi-square test, so I use it in the new function.

[xuexingwangzhedishitian]

thank you for your reply.
I have been read the CNV analysis workflown(https://docs.gdc.cancer.gov/Data/Bioinformatics_Pipelines/CNV_Pipeline/).
I still have some problems.
The genotype files with continuous copy number values are generated by GITSIC. Based on the cutoff of 0.3, copy number values could be transfered into -1,0 and 1. Processed file is named the Gene Level Copy Number Scores files. I want to know the funciton of differential_cnv is applicable for the output genotype files with continuous copy number values generated by GITSIC.
I am glad for you reply.

[huerqiang]

It is not suitable for continuous data and needs to be converted to discrete data.

[xuexingwangzhedishitian]

I mean wether the 0.5 cutoff of the function of differential_cnv is applicable to the output focal-level files that generated by GITSIC as follows.
Gene Symbol Gene ID Cytoband TCGA-27-2519-01A-01D-0784-01 TCGA-06-5408-01A-01D-1694-01
ACAP3 116983 1p36.33 0.000 -0.873
ACTRT2 140625 1p36.32 0.000 -0.873
.....
Thank you for your reply again.

[huerqiang]

I suggest that the threshold value should be 0.3 recommended by TCGA official website now.

[xuexingwangzhedishitian]

Thanks for your reply. I get it.