https://github.com/varfish-org/hgvs-rs
A port of biocommons/hgvs to the Rust programming language
https://github.com/varfish-org/hgvs-rs
genetics varfish variant-filtration variant-prioritization vcf
Last synced: about 1 month ago
JSON representation
A port of biocommons/hgvs to the Rust programming language
- Host: GitHub
- URL: https://github.com/varfish-org/hgvs-rs
- Owner: varfish-org
- License: apache-2.0
- Created: 2023-02-13T18:11:24.000Z (about 3 years ago)
- Default Branch: main
- Last Pushed: 2026-04-01T10:40:56.000Z (about 1 month ago)
- Last Synced: 2026-04-01T12:25:21.617Z (about 1 month ago)
- Topics: genetics, varfish, variant-filtration, variant-prioritization, vcf
- Language: Rust
- Homepage:
- Size: 1.33 MB
- Stars: 17
- Watchers: 4
- Forks: 6
- Open Issues: 19
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- License: LICENSE.txt
Awesome Lists containing this project
README
[](https://crates.io/crates/hgvs)
[](https://crates.io/crates/hgvs)
[](https://crates.io/crates/hgvs)
[](https://github.com/varfish-org/hgvs-rs/actions/workflows/rust.yml)
[](https://codecov.io/gh/varfish-org/hgvs-rs)
[](https://zenodo.org/badge/latestdoi/601272076)
# hgvs-rs
This is a port of [biocommons/hgvs](https://github.com/biocommons/hgvs) to the Rust programming language.
The `data::cdot::*` code is based on a port of [SACGF/cdot](https://github.com/SACGF/cdot) to Rust.
## Choosing a Provider
The crate includes three `Provider` implementations for accessing transcript and sequence data:
| Provider | Data Source | Best For |
|---|---|---|
| `data::uta::Provider` | PostgreSQL (UTA) | Testing, development |
| `data::uta_sr::Provider` | PostgreSQL (UTA) + SeqRepo | Testing with genome contigs |
| `data::cdot::json::Provider` | cDOT JSON + SeqRepo | Offline use without PostgreSQL |
All three providers incur I/O overhead per query (database round-trips or filesystem access), which
can dominate normalization throughput for high-volume workloads.
For **production use**, consider implementing the `Provider` trait with an in-memory or
purpose-built data store. Two examples of this approach:
- [mehari](https://github.com/varfish-org/mehari) deserializes a protobuf-encoded transcript
database into memory and indexes it with interval trees, backed by RocksDB for sequence storage.
- [ferro-hgvs](https://github.com/fulcrumgenomics/ferro-hgvs) uses indexed FASTA files with
cDOT metadata, supporting random access via FAI indexes and optional memory-mapped I/O.
## Running Tests
The tests need an instance of UTA to run.
Either you setup a local copy (with minimal dataset in `tests/data/data/*.pgd.gz`) or use the public one.
You will have to set the environment variables `TEST_UTA_DATABASE_URL` and `TEST_UTA_DATABASE_SCHEMA` appropriately.
To use the public database:
```
export TEST_UTA_DATABASE_URL=postgres://anonymous:anonymous@uta.biocommons.org:/uta
export TEST_UTA_DATABASE_SCHEMA=uta_20210129
```
Note that [seqrepo-rs](https://github.com/varfish-org/seqrepo-rs) is used for access to the genome contig sequence.
It is inconvenient to provide sub sets of sequences in SeqRepo format.
Instead, we use a build-cache/read-cache approach that is also used by `biocommons/hgvs`.
To build the cache, you will first need a download of the seqrepo [as described in biocommons/biocommons.seqrepo Quickstart](https://github.com/biocommons/biocommons.seqrepo#quick-start).
Then, you configure the running of tests for `hgvs-rs` as follows:
```
export TEST_SEQREPO_CACHE_MODE=write
export TEST_SEQREPO_PATH=path/to/seqrepo/instance
export TEST_SEQREPO_CACHE_PATH=tests/data/seqrepo_cache.fasta
```
When running the tests with `cargo test`, the cache file will be (re-)written.
Note that you have to use `cargo test --release -- --test-threads 1 --include-ignored` when writing the cache for enforcing a single test writing to the cache at any time.
If you don't want to regenerate the cache then you can use the following settings.
With these settings, the cache will only be read.
```
export TEST_SEQREPO_CACHE_MODE=read
export TEST_SEQREPO_CACHE_PATH=tests/data/seqrepo_cache.fasta
```
After either this, you can run the tests.
```
cargo test
```
## Creating Reduced UTA Databases
The script `tests/data/data/bootstrap.sh` allows to easily build a reduced set of the UTA database given a list of genes.
The process is as follows:
1. You edit `bootstrap.sh` to include the HGNC gene symbols of the transcripts that you want to use.
2. You run the bootstrap script.
This will download the given UTA dump and reduce it to the information related to these transcripts.
```
$ bootstrap.sh http://dl.biocommons.org/uta uta_20210129
```
The `*.pgd.gz` file is added to the Git repository via `git-lfs` and in CI, this minimal database will be used.
## Some Timing Results
(I don't want to call it "benchmarks" yet.)
### Deserialization of large cdot JSON files.
Host:
- CPU: Intel(R) Xeon(R) E-2174G CPU @ 3.80GHz
- Disk: NVME (WDC CL SN720 SDAQNTW-1T00-2000)
Single Running Time Results (no repetitions/warm start etc.)
- ENSEMBL: 37s
- RefSeq: 67s
This includes loading and deserialization of the records only.