Description

DEPF framework

The DEPF package contains four files: ./scRNAseqdata, ./HierarchicalAutoencoder, ./BiobjectiveFruitFlyOptimizationAlgorithm, and ./OutputData.

  • ./scRNAseqdata: Storing raw scRNA-seq data sets

  • ./HierarchicalAutoencoder: Storing Normalization module, Hierarchical Autoencoder module, Louvain clustering module, Leiden clustering module.

  • ./BiobjectiveFruitFlyOptimizationAlgorithm: Storing Spectral clustering module, dynamic ensemble pruning mudule.

  • ./OutputData: Storing latent data, Louvain clustering results, Leiden clustering results, final clustering results, etc.

The operational process of DEPF involves three parts: RunHA.R, runLouvain.R || runLeiden.R || runSpectral.m, and runBioFOA.m.

  • RunHA.R: Normalizing and mapping the raw -scRNA-seq data to a low-dimensional latent space.

  • runLouvain.R || runLeiden.R || runSpectral.m: Selecting a basic clustering algorithm to generate a clustering ensemble.

  • runBioFOA.m: Performing dynamic ensemble pruning.