paper · benchmark creator

Crafted experiments to evaluate feature selection methods for single-cell RNA-seq data

Lineberger Comprehensive Cancer Center, University of North Carolina · University of North Carolina at Chapel Hill · 2025-03-19

Relationship layer

Benchmark usage

This table records what the work did with each benchmark before attempting to normalize a run. Partial claims remain visible without being treated as comparable evaluations.

benchmark creation

crafted-experiments-to-evaluate-feature-selection-meth-crafted-experiments-1-use

Non-evaluationunknown

Benchmark: crafted experiments · version Not reported

Selection
not applicable
Models
Not reported / not applicable
Metrics
Not reported / not applicable
Linked runs
None

Not reported / unresolved: Benchmark version is not reported.; An explicit benchmark release date is not reported.; The released crafted-dataset artifact's license is not reported.

AI-assisted double-pass extraction; values are limited to independently supported claims.

Evidence
  • section: Introduction
    Supports: /relation_type
  • section: Introduction
    Supports: /benchmark_id

evaluation

crafted-experiments-to-evaluate-feature-selection-meth-crafted-experiments-2-use

Partialunknown · n=24

Benchmark: crafted experiments · version Not reported

Selection
not reported · Guided tutorials and default parameters; top 2000 ranked features per method, except HIPPO used its zero-proportion-test cutoff
Metrics
MultiK low-resolution rank, DiProPerm Z scores, average rank, proportion of crafted genes selected, number of crafted genes selected, number of crafted genes not selected
Linked runs
None

Not reported / unresolved: Benchmark version is not reported.; Exact software versions and random seeds are not reported.; Figure 4 bar values are unlabeled, so no numeric model results were extracted.; Conflicted model claim omitted after independent verification; the evaluation relationship is published conservatively.; Conflicted repeats claim omitted after independent verification; the evaluation relationship is published conservatively.; benchmark version; numeric result

AI-assisted double-pass extraction; values are limited to independently supported claims.

Evidence
  • section: Results — Crafted experiment applications
    Supports: /relation_type
  • section: Results — Crafted experiment applications
    Supports: /benchmark_id
  • section: Results — Crafted experiment applications
    Supports: /scope
  • section: Results — Crafted experiment applications
    Supports: /scope
  • section: Results — Crafted experiment applications
    Supports: /scope
  • section: Materials and methods — Benchmarking of feature selection methods
    Supports: /scope
  • section: Results — Crafted experiment applications; Reference 16
    Supports: /model_ids
  • section: Results — Crafted experiment applications; Reference 15
    Supports: /model_ids
  • section: Results — Crafted experiment applications; Reference 18
    Supports: /model_ids
  • section: Results — Crafted experiment applications
    Supports: /model_ids
  • other: Article author list; Results — Crafted experiment applications
    Supports: /model_ids
  • other: Article author list; Results — Crafted experiment applications
    Supports: /model_ids
  • other: Article author list; Results — Crafted experiment applications
    Supports: /model_ids
  • section: Materials and methods — Benchmarking of feature selection methods
    Supports: /metric_labels
  • section: Materials and methods — Benchmarking of feature selection methods
    Supports: /metric_labels
  • section: Results — Crafted experiment applications
    Supports: /metric_labels
  • section: Results — Crafted experiment applications
    Supports: /metric_labels
  • section: Results — Crafted experiment applications
    Supports: /metric_labels
  • section: Results — Crafted experiment applications
    Supports: /metric_labels

Normalized evaluation runs

This source has no normalized model run. It may be a creator-only source or a partial/non-evaluation benchmark use.