Post-proces SURVIVOR output with vcf2bedpe.R

[arnikz]

In this workflow, vcf2bedpe.R is only used to convert SV callers' output. According to GooglingTheCancerGenome/sv-gen#43, we also need to post-process SURVIVOR simSV output but this script seems missing in the workflow.

[lsantuari]

I will update the workflow (related to #68).

[arnikz]

StructuralVariantAnnotation (SVA) package supports the following VCF notations:

• Non-symbolic allele
• Symbolic allele with SVTYPE of DEL, INS, and DUP.
• Breakpoint notation SVTYPE=BND
• Single breakend notation

Currently, we use the two notations (BND for TRA/CTX?) in the workflow.

[arnikz]

SVA supports non-compliant VCFs:

• Pindel (SVTYPE=RPL)
• Manta (INv3, INV5 fields)
• Delly (SVTYPE=TRA, CHR2, CT fields)
• TIGRA (SVTYPE=CTX)

Is it correct to say that vcf2bedpe.R script (and bedpe2vcf.py) is needed to handle also LUMPY and SURVIVOR output?

[lsantuari]

StructuralVariantAnnotation (SVA) package supports the following VCF notations:

• Non-symbolic allele
• Symbolic allele with SVTYPE of DEL, INS, and DUP.
• Breakpoint notation SVTYPE=BND
• Single breakend notation

Currently, we use the two notations (BND for TRA/CTX?) in the workflow.

Yes. I miss the inversions (INV) in the SVA documentation.

Based on my tests SVA supports also INV (see page 3 of VCF specs).
INV can also be represented in BND notation (example to be added here based on this).

[arnikz]

Yes. I miss the inversions (INV) in the SVA documentation.

True. Nevertheless, INVs are handled, for example, in vcf2bedpe.R, label_classes.py and/or notebook [8].

[arnikz]

The script is used to process Manta, DELLY, LUMPY, GRIDSS and SURVIVOR VCF files. However, the following lines do not make it clear for the latter (e.g., TRA vs. CTX).

sv-channels/scripts/R/vcf2bedpe.R

Line 14 in 50b1a55

arg_parser("Convert VCF output of Manta, DELLY, LUMPY or GRIDSS to BEDPE format.")

sv-channels/scripts/R/vcf2bedpe.R

Line 52 in 50b1a55

"CTX",

sv-channels/scripts/R/vcf2bedpe.R

Line 71 in 50b1a55

# if the file was generated with SURVIVOR simSV, treat translocations as DELLY TRA entries

sv-channels/scripts/R/vcf2bedpe.R

Lines 95 to 96 in 50b1a55

	# TRA
	idx <- which(info(sv_callset_vcf)$SVTYPE == 'TRA')

Moreover, all issues reported here should be handled by this script. Correct?