Support for CNVs

[pasted]

VCF representation of a CNV can be difficult to handle, wit non-specific end points and variable quality of frequency data (DGV http://dgv.tcag.ca/dgv/app/home).

CNV HTZ / HMZ deletion model should:

• return LOF details for all genes that overlap region
• Aggregate stats on numbers of HMZ / HTZ LOF variants within regions on a gene by gene basis
• if high number of LOF variants from Gnomad in deleted region = lower pathogenicity
• Given Gnomad data covers regions outside of Exonic boundaries should probably also look at intergenic regions if covered by CNV

CNV duplication model should:

• return the same details as deletion CNV but with different conclusions