Welcome to MuPPI Dataset documentation!

MuPPI Dataset

In this multi-view dataset we describe the proteins of a protein-protein interaction (PPI) network. The aim is to classify the extremely multifunctional* (EMF) proteins of others, and to search for their possible characteristics. The PPI network and the list of EMFs are taken from the work of Charles Chapple and Christine Brun*, and are available from the MoonDB** database. The PPI network is created from several different databases, with filtering at the level of interaction identification methods. The MoonGO* software considers only direct, binary and experimentally tested interactions. Interactions of a protein with itself are suppressed. This network is composed of 13683 proteins, and 90386 interactions.

Documentation

Release

0.0.0

Date

Aug 11, 2023

Contents:

• Readme
• Multiview Dataset MuPPI
  • Labels:
  • PPI network topology view:
  • Subcell location view:
  • Tissue Expression view:
  • PPInetwork Embedding view:
  • Gene Ontology view:
  • GO PPInetwork Embedding view:
  • Phenotype Ontology view:
  • Protein Domains view:
  • Post Translational Modifications view:
  • 3’UTR Complexes view:
  • Linear Motifs view:
  • Cancer Mutations view:
  • References
• Benchmark of MuPPI Dataset
  • Introduction Benchmark of MuPPI Dataset
    • Algorithms
    • Learning Tasks
    • Results summary
  • Results of Benchmak with Raw Data
    • First task : ‘EMF’ versus mono-clustered
    • Second task : multi_clustered vs mono_clustered
    • Last task : EMF versus multi-clustered
  • Sub-sampling
    • Accuracy
    • F1-score
    • Comparison with un-balanced data
    • Features Importance ??
    • Samples analysis ??
  • Over-sampling
    • Method
    • Accuracy
    • F1-score
    • Comparison with un-balanced data

Indices and tables

• Index

• Module Index

• Search Page