Processing Whole Genome Sequencing data (shallow) and call Copy Number Variations

Question

[Alejandro Jimenez Sanchez]

Dear all,

We are proposing a project where we want to perform shallow Whole Genome Sequencing (<1x) on DNA obtained from tumour FFPE human samples.

I have worked before with Whole Exome Sequencing and I have called SNVs and so on, but I haven't worked with WGS yet.

Does anyone know of any good tools and steps to get copy number variations from shallow WGS?

I haven't seen any paper that benchmarks tools yet, but if someone knows about this topic and could give some light in this, I would appreciate it very much.

Answer 1

[AlexRT]

If it's for germline copy-number variations, the 1000 genomes projects has called structural variations including copy number changes from individuals with ~7X WGS (https://www.nature.com/articles/nature15394). You can use GenomeStrip (http://software.broadinstitute.org/software/genomestrip/genome-strip-20) or Delly (https://github.com/dellytools/delly), or Lumpy (https://github.com/arq5x/lumpy-sv).

Since you have really low-depth WGS, the key is to have enough samples to pick up power for detecting variants in carriers vs non-carriers in the population. I never used <1x WGS for copy-number, but it might be harder to detect call copy-neutral variants such as inversions.

[Alejandro Jimenez Sanchez]

It is for copy number variations in tumour tissues, so germline changes would be removed.

[AlexRT]

Oups forget the comment then.