A Bioinformatician's Storyboard (Notebook Idea)

[wesleygoar]

A bioinformatician wants to update the sequencing annotation and analysis pipeline at their institution. They inherited the pipeline from someone else and found a lot of bespoke code to allow for variant annotation from disparate resources. The majority of data they are using for annotation is now stored on their local machine, cluster, or cloud instance. These data are stored in many different formats but for the sake of this storyboard, all data is on the same reference build.

Nomenclature

• HGVS
• gnomad style VCF
• VCF (chrom, position, ref, alt separate fields)
• Free text

The bioinformatician would love to streamline the annotation process and have greater confidence in the accuracy of translation between different formats such as VCF & HGVS (different alignment strategies) for annotation since the annotated data is used for clinical decision making. They hear about the good word of VRS and decided they would like to try using VRS within their pipeline to act as the key identifier between different data sources. They ask the following questions to themself:

1. How can I annotate my VCF file with VRS so I can use it as the key to annotate my VCF with all my other annotation data? How modular is the annotator? (What options do I have when performing annotation?)
2. How can I add VRS to all my different data sources so I can use it as my lookup key?

What are the key operations I need to know to do this?
Will VRS work for all my variants? If not, how does that affect my workflow?

3. Are there other data sources that natively store VRS in their data structure? If so, how do I ensure I am using the same version of VRS as they are?

[wesleygoar]

Add in our experience adding VRS into gnomAD using the vcf_annotator as a use case.

• Talk about technical aspects
• reference build
• VCF spec
• example VCF being annotated

[wesleygoar]

Can take pieces of what we put into the gnomAD blost post