Beagle 4.0 Imputation - Warning: Skipping VCF record with no data

Question

[Wietje]

Hello fellow Biostars,

I am currently working on an imputation from 6k to 50k data and one of the programmes in use is Beagle 4.0 (since I want to integrate pedigree data). When I performed a test run on a single chromosome, the following warning came up for 456 out of 986 SNPs:

"Skipping VCF record with no data" ---> Do you have any idea how this can happen?

I checked the manual/ documentation and other platforms and couldn't find any other comparable reports. Interestingly, the imputation was performed nevertheless and the accuracy seems to be very good as well. Still, I would like to get to the bottom of this problem/ warning.

All help and ideas are heartily appreciated!

Answer 1

[Pierre Lindenbaum]

called here: https://github.com/tfwillems/PhasedBEAGLE/blob/master/vcf/VcfWindow.java#L341 when e.isMissingData()

defined here: https://github.com/tfwillems/PhasedBEAGLE/blob/master/vcf/VcfEmission.java#L70

Returns 'true' if the genotype emission probabilities for each sample are determined by a genotype that contains only missing alleles, and returns 'false' otherwise. Return 'true' if the genotype emission probabilities for each sample are determined by a genotype that contains only missing alleles, and 'false' otherwise.