Data

Tools

Indexes

Applications

Experiments

Open Source Science

https://github.com/bblodfon/pdac-efs-bench2024

Feature Selection benchmark and analysis on multi-omics PDAC datasets

https://github.com/bblodfon/pdac-efs-bench2024

Science Score: 49.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ✓
    .zenodo.json file
    Found .zenodo.json file
  • ✓
    DOI references
    Found 3 DOI reference(s) in README
  • ✓
    Academic publication links
    Links to: arxiv.org, zenodo.org
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (11.3%) to scientific vocabulary
Last synced: 6 months ago · JSON representation

Repository

Feature Selection benchmark and analysis on multi-omics PDAC datasets

Basic Info
  • Host: GitHub
  • Owner: bblodfon
  • License: mit
  • Language: R
  • Default Branch: main
  • Homepage:
  • Size: 95 MB
Statistics
  • Stars: 1
  • Watchers: 1
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created over 1 year ago · Last pushed 6 months ago
Metadata Files
Readme License

README.md

pdac-efs-bench2024

DOI

This repository contains code and data for benchmarking multi-omics feature selection methods in pancreatic ductal adenocarcinoma (PDAC), with a focus on survival prediction.

Citation

Zobolas, J., George, A.-M., López, A., Fischer, S., Becker, M., & Aittokallio, T. (2025). Optimizing Prognostic Biomarker Discovery in Pancreatic Cancer Through Hybrid Ensemble Feature Selection and Multi-Omics Data. https://arxiv.org/pdf/2509.02648

Environment Setup

Restore R libraries required to run the benchmark using renv: renv::restore(exclude = c("BiocGenerics", "BiocManager", "BiocVersion", "ComplexHeatmap", "IRanges", "S4Vectors", "MASS", "ipred", "class"))

Data and Preprocessing

We use 3 multi-omics PDAC datasets. Preprocessing steps include:

  • Patient and feature filtering
  • Metadata curation
  • Resampling definition for benchmark
  • Creation of mlr3 survival tasks per omic data type

All related code and metadata are found in the data/ directory.

Multi-omics Benchmark Overview

The benchmarking pipeline consists of three main steps, located in the bench/ directory.

Hybrid Ensemble Feature Selection (hEFS)

Script: bench/run_efs.sh: Runs the Ensemble Feature Selection procedure (bench/efs.R) across:

  • All datasets
  • All omics
  • All resampling iterations

Stores results as EnsembleFSResult objects in bench/efs/. Contact the main author to share with you these intermediate results.

Omic-wise Feature Selection

Script: bench/run_fs.sh: Performs feature selection per omic and per subsampling iteration (100 total).

Two methods available: - Cox Lasso - Pre-computed Ensemble Feature Selection (loaded from step 1). We automatically select the number of features via the Pareto front method.

Output: bench/fs.rds - a table with:

  • dataset_id, omic_id, rsmp_id
  • Selected features for each method

Multi-omics Integration & Benchmarking

Script: bench/run_mm_bench.R: Combines selected features across omics (via late integration/fusion) per subsampling iteration, then trains and evaluates survival models on training/test splits. Available models are Cox Proportional Hazards, Cox Lasso, Random Survival Forests and BlockForest.

Output: bench/result.rds - a table with:

  • dataset_id: Identifier of the dataset used
  • fs_method_id: Feature selection method applied
  • rsmp_id: Subsampling (resampling) iteration identifier
  • model_data_config: Configuration used for model training, indicating the model type and which omics and/or clinical data were included (all means clinical + all omics)
  • task_nfeats: Number of selected features used in the task
  • task_feats: The specific features selected
  • Performance scores for the test sets (Harrell's C-index, etc.)

Owner

  • Name: John Zobolas
  • Login: bblodfon
  • Kind: user

GitHub Events

Total
  • Watch event: 1
  • Member event: 1
  • Push event: 118
  • Public event: 1
Last Year
  • Watch event: 1
  • Member event: 1
  • Push event: 118
  • Public event: 1

Issues and Pull Requests

Last synced: 11 months ago

All Time
  • Total issues: 0
  • Total pull requests: 0
  • Average time to close issues: N/A
  • Average time to close pull requests: N/A
  • Total issue authors: 0
  • Total pull request authors: 0
  • Average comments per issue: 0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
Past Year
  • Issues: 0
  • Pull requests: 0
  • Average time to close issues: N/A
  • Average time to close pull requests: N/A
  • Issue authors: 0
  • Pull request authors: 0
  • Average comments per issue: 0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
View more stats
Top Authors
Issue Authors
Pull Request Authors
Top Labels
Issue Labels
Pull Request Labels