GeneBe Hub RFCs published

[pstawinski]

Request for Comments: GeneBe Hub

TLDR;

Public, open repository for variant annotation databases is suggested, named GeneBe Hub. One can easily annotate VCF file with prepared databases. Repositry is open, you can publish your annotations there. There are 3 elements that builds the system:

1. Most important: the unified Format for databases
2. The Client to apply the annotations
3. The Hub to browse published databases.

We publish three RFC to discuss proposition, while the most important is the Format.

---

Summary

GeneBe Hub is an open, public repository of Genetic Variant Annotation Databases. It aims to provide a unified platform for sharing, discovering, and utilizing genetic variant annotations. By introducing a standardized format, repository and a client application for processing these annotations, GeneBe Hub addresses the challenges of inconsistency, lack of centralization, and complex application methods.

We are inviting the bioinformatics community to provide feedback on the proposed format, distribution methods, and application workflows to ensure GeneBe meets the needs of researchers, geneticists, and developers alike.

The first round of feedback collection will conclude on 21th February 2025.

---

Background

Challenges in Genetic Variant Annotation

Interpreting genetic variants requires extensive data. Before assigning pathogenicity to a variant, an expert must consider various factors, such as:

• Population frequencies (e.g., GnomAD, ALFA).
• Variant effects and gene impact (e.g., SNPeff, VEP).
• Associations with phenotypes (e.g., ClinVar).
• Specialized datasets providing population-specific frequencies or computational scores for specific genes.

While numerous tools and databases address these aspects, the field lacks a universal standard for sharing, applying, and distributing variant annotations. Databases are often distributed in diverse formats (VCF, TSV, bigWig), making it challenging to parse and integrate them. Additionally, these databases are published in various locations, often making them difficult to find and explore. With continuously updated resources like ClinVar, keeping up-to-date with the latest data becomes increasingly complex for bioinformaticians.

Variant Annotation Distribution Challenge

There are hundreds of variant annotations available. If you are the author of a new variant scorer, frequency aggregator, or another type of variant annotation, getting it to users is not easy. You can publish your annotations on your website, submit a paper to a scientific journal, and... wait. But what format should you use? There is no standard yet. How can you make your annotation searchable and applicable?

You could develop an application that applies your annotations to VCF files, but will users actually download and use it? This challenge prevents many great annotations from gaining wider recognition—simply because they are difficult to discover and apply.

---

The GeneBe Solution

• We want to let clinicians and bioinformatities to easily find new annotations by browsing single website, and then to apply annotations to their data just by writing calling a local program like: annotate my.vcf annotator-name:version
• We want to give annotation creators a tool for easy creating and distributing new annotations using the language and framework they already know and use.

GeneBe Hub aims to solve these challenges by offering:

1. A Standardized Format

   • A unified format for storing and sharing annotations.
   • Documentation and examples available at: GeneBe Hub Format.

2. A Client Application

   • GeneBe Client allows users to:
     • Download and apply annotations to VCF files.
     • Create and share new annotation databases.
   • Available at: GeneBe Client.

3. A Centralized Repository

   • A platform for users to discover, share, and distribute annotations.
   • Visit: GeneBe Hub.

---

Non-Goals

GeneBe Hub does not aim to:

• Develop another variant effect predictor: Tools like SNPeff and VEP already excel in this area. GeneBe focuses on facilitating the sharing and application of annotations, not predicting variant effects.

---

Key Goals of this RFC

We are seeking community input on the following aspects:

1. Standardized Format

• We present a format for variant annotations, applied by variant change, variant position or position range. We would like to discuss the pros and cons of the solution.
• We start a disussion about a format for gene and transcript annotations.
• We present possible extensions to the format.

Please continue reading on Database Format RFC

2. Distribution via (not only) GeneBe Hub

• We present the GeneBe Hub, a solution for easy sharing of the databases. We would like to present the motivation and the solution and hear the feedback.

Please continue reading on Public repository RFC

3. Application Methods / GeneBe Client

• We have prepared a simple client application for annotating VCF files and for working with databases. We would like to discuss the strengths and limitations of the solution as well as discuss the possible development directions.

Please continue reading on GeneBe Client RFC

---

Feedback

We encourage all interested parties to:

1. Participate in the discussion on GitHub for each of the topics: GeneBe Hub RFC.

Promotion

This RFC will also be promoted on Reddit, Biostars and other platforms to gather diverse perspectives. We value input from geneticists, bioinformaticians, software developers, and anyone engaged in the interpretation of genetic data. We would also be grateful for help in spreading the RFC.

---

Quick links:

• Website: GeneBe Hub
• Documentation: GeneBe Format
• Client Application: GeneBe CLI
• Discussion on the RFCs: RFCs
• Format RFC
• Client RFC
• Hub RFC

[pstawinski]

First Round of the RFC is Over – Thank You!

The first round of the RFC has officially concluded. A huge thank you to everyone who shared their ideas via private messages! Your input has been greatly appreciated.

While this round is over, the discussion remains open—if you still have suggestions, feel free to send them.

A summary of the first round will be shared soon. Stay tuned!

[pstawinski]

Conclusions so far:

1. There is a need for a more complex annotation key. We need to support: (1) variant + transcript, (2) transcript + position, and (3) transcript + amino acid change-keyed annotations.
2. We need to support lists of primitives and lists of tuples of primitives as annotation columns.
3. In some cases, partitioning Parquet by _seq may be problematic. It should be possible to have databases that are not partitioned.

[pstawinski]

In order to be able to support many platforms to read GeneBeHub stored data there must be limits on the names of the fields and the database itself.

The biggest common supported set of safe characters are [A-Za-z0-9_] , name must start with [A-Za-z] and there must be no following two or more _ and name cannot end with _.