Exclusion Of Regions In Cnv Analysis

Question

[bob.obr13n]

Hi,

In various CNV analysis papers there is mention of excluding areas of the Genome from analysis

e.g. if the CNV spanned a centromere, if the CNVs overlapped at least 50% with regions previously defined as being prone to false positives due to somatic mutations, etc

Would anyone have a paper that details a list of which regions to exclude from CNV analysis, or a preprepared file for PLINK?

Best regards

Bob

Answer 1

[Malachi Griffith]

There are several tracks in UCSC that might be useful for selecting regions to exclude from CNV analysis. For example if you wanted to exclude near centromeres and telomeres, regions of high or low GC, assembly gaps, regions with poor mappability, repeat regions, etc. Here are some examples of such tracks with links to their descriptions

• Chromosome Band
• Gap Locations
• GC Percent
• GRC Incident
• HiSeq Depth
• Mappability
• Repeat Masker

[wanabi]

Thanks for the info.

What mappability threshold would you use? Would you use Alignability or Uniqness definitions?

Thanks!

[NS]

Hello, Could you please help me how to exclude highly polymporphic regions and Chr X, Y regions. I am new in the NGS data. So please guide. How to get its bed file

Answer 2

[Vikas Bansal]

Hi! Could you mention the papers ("In various CNV analysis papers")? In my analysis, I have excluded repetitive regions (satellite, SINE, LINE, etc). You can download this file from UCSC (repeat masker, rmsk). Can you please mention the species?. Also, I have excluded common copy number alterations present in DGV - my analysis is to find rare disease causing copy number alterations (I assumed that you are talking about human genome) but you should always double check (manually by looking at excluded regions from DGV) because some of regions filtered out by DGV can be interesting for you.