Suitability of OpenCRAVAT's Reporting Interface for Population-Wide Variant Characterization

[arakelyanaa]

Hi,

We have launched a project focused on population-wide variant characterization in the Armenian population. As part of this effort, we are seeking a suitable interface to report aggregate-level results, including variant frequencies, annotations, and comparisons with gnomAD populations.

We are considering OpenCRAVAT due to its comprehensive variant annotation and reporting capabilities. However, we have a few questions regarding its suitability for our project:

1. How suitable is OpenCRAVAT’s reporting interface for population-wide variant characterization? Specifically, are there any existing recommendations or tutorials that guide users in utilizing OpenCRAVAT for this purpose?

2. We have noticed that sample names are displayed in the Variant tab, which raises ethical and privacy concerns. Could you provide guidance on modifying the reporting interface to hide or remove the sample names column to protect individual identities?

3. Can single variant reporting be implemented locally? If so, are there any best practices or recommendations for setting this up?
   Thank you, and I look forward to your feedback.

Best regards,
Arsen

[RachelKarchin]

Thanks for your interest! Please send us more detail about your use case, such as the size of the population you are working with and the variant characteristics that you want to identify. In the meantime, here are some answers to your initial questions. We are also glad to arrange a meeting with you!

How suitable is OpenCRAVAT’s reporting interface for population-wide variant characterization?

Do you want to characterize variants from a very large number of VCF or genotype files? We can take very large numbers of VCF files as input.

Specifically, are there any existing recommendations or tutorials that guide users in utilizing OpenCRAVAT for this purpose?

Not specifically but we would be glad to put a tutorial together for you.

We have noticed that sample names are displayed in the Variant tab, which raises ethical and privacy concerns. Could you provide guidance on modifying the reporting interface to hide or remove the sample names column to protect individual identities?

Yes. This should be straightforward.

Can single variant reporting be implemented locally? If so, are there any best practices or recommendations for setting this up?

Yes. This can be done by running OpenCRAVAT on the command line as follows:
oc module install variantreport
This will install a large number of annotators and their associated data as dependencies and requires approximately 500GB of disk space.

[arakelyanaa]

Hi, thanks for the reply!
We plan to do whole exome sequencing of 1k-2k people and obtain up to 500 whole genomes in the next two years. However, the project will be continuous. We do variant calling according to the GATK4 Best Practices for SNVs and small indels. Now we are generating gVCF files per run and then merging them. We plan to use OpenCravat for the population-wide frequencies of identified variants and annotate them with selected Variant, Frequency, and Clinical Relevance Annotators. As an output, we expect to get summary-level information about the variants in the population. What concerns the Single Variant Reporting, is it possible to use a selected set of variants, and use data-specific information (taken from our VCFs) in the report? We run OpenCravat locally, on a small server with 16Gb RAM and 4Tb storage in a docker container, but we are considering moving it to a larger one.
It would be great if we could schedule a call at your earliest convenience to discuss these issues.

Best regards,
Arsen

[RachelKarchin]

A call sounds great! Please send your email to support@opencravat.org and we will reach out to set up a meeting time.