Inside Genebench-Pro

Case studies

These 10 case studies showcase representative questions from GeneBench-Pro. Each case study includes the original prompt, datasets, and supporting materials. For an overview of the benchmark and key findings, see the announcement blog.

Note: File previews show excerpts from the full datasets.

Case study 1

Somatic oncology: Structural variant-guided tumor therapy benefit-risk decision

Estimate whether a synthetic TXR1-directed inhibitor has positive clinical utility in tumors whose target activation is driven by a structural variant. TXR1, TXR1i, DLR1, and star-allele labels are synthetic benchmark labels. 

The target subgroup has to be recovered from long-read, expression, tumor-quality, and pharmacogenomic evidence before benefit and toxicity can be interpreted as a treatment decision.

Files provided to the model

Case study 2

Functional genomics: CRISPR target validation: lncRNA transcript or genomic locus?

Decide whether an apparent lncRNA dependency is transcript-specific or driven by nearby-locus and neighbor-gene effects.

Transcript-directed evidence has to survive controls for local DNA-locus perturbation, neighbor-gene repression, guide swaps, GC toxicity, and plate effects.

Files provided to the model

Case study 3

Statistical genetics: Prioritizing protein drug targets in a linked genetic locus

Estimate direct disease effects for two nearby proteins using cis multivariable Mendelian randomization (cis-MVMR) while handling assay scale, allele orientation, winner's curse, LD, and residual local pleiotropy.

The two proteins share a correlated locus. The analysis has to move from marginal associations to conditional, LD-aware disease effects on a common protein scale.

Files provided to the model

Case study 4

Clinical genomics / carrier screening: DRX1 carrier-screening residual risk under CNV and pseudogene calibration

Files provided to the model

Case study 5

Single-cell genomics: Activated-monocyte eQTL after ambient RNA correction

Files provided to the model

Case study 6