Main Content

How are the five categories of copy number changes determined?

Submitted by gaheens on

The GDC begins with integer-level estimates of absolute copy number generated by either the ASCAT or ABSOLUTE pipeline. To establish a baseline, an integer-valued sample ploidy is computed as follows: 

  • For gene-level CNV, the mode of copy number values is used across all autosomal protein-coding genes
  • For segment-level CNV, a length-weighted mode of copy number values is computed across all autosomal segments
  • In cases of a tie, the mode is rounded up
  • Please note that the integer-valued sample ploidy used here differs from the floating-point ploidy estimates produced directly by the ASCAT or ABSOLUTE pipelines. The latter should be considered the more precise representation and is recommended for use in most other bioinformatics analyses.

Based on this sample ploidy value, the GDC assigns copy number categories as: 

  • Homozygous deletion: copy number = 0 Loss: 0 < copy number < sample ploidy 
  • Neutral: copy number = sample ploidy 
  • Gain: sample ploidy < copy number < 2 × sample ploidy 
  • Amplification: copy number ≥ 2 × sample ploidy
Subject Tag