Performance Metrics for SPiP Benchmark Dataset

Performance metrics for six benchmark datasets used to evaluate SpliceAI and its open-source implementations. The data includes results from 1,316 validated variants, 213 variants with splice-assay data, 99,601 variants from the SPiP study, 242 deep intronic pathogenic variants, and two ClinVar-derived datasets totaling over 111,000 variants. Authored by Nathan Fortier and last updated in May 2026, the dataset is shared under a CC-BY-4.0 license.

Use Cases

• Benchmarking splice prediction algorithms based on performance metrics across multiple variant datasets.
• Calibrating score thresholds for deep intronic variant analysis based on reported optimal thresholds.
• Comparing the concordance of open-source SpliceAI implementations against the original tool.
• Evaluating algorithm performance on canonical versus non-canonical splice site variants.

Strengths

• Includes performance results across six distinct datasets, providing a multi-context evaluation.
• Contains results for over 213,000 variants in total, including a large set of 99,601 variants from the SPiP study.
• Reports specific metrics like balanced accuracy (0.841-0.977) and splice-site match rates exceeding 90%.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale modeling.

Provenance

Source

figshare

Collection Method

Likely contains aggregated performance metrics from comparative benchmarking studies.

Freshness

Last updated 2026-05-13 17:43:01; freshness should be verified.