Hello everyone. We are trying to run the sarek pipeline with our local VEP cache using Illumina's whole genome sequencing data and we encounter an issue where VEP cannot annotate the multiallelic sites, therefore the variant-called VCF has to be normalized to be able to successfully annotate the variants. Is there a convenient way of normalizing VCF before giving it to VEP?

you can try

--normalize_vcfs

https://nf-co.re/sarek/3.7.1/parameters/#normalize_vcfs

that's exactly what I need, thanks!

If you already ran the pipeline, you can just redo the relevant parts using the

--step

parameter. And if you are using an older version of sarek (without built-in normalization), you can manually normalize and then run with

--step annotate

.