HEMDAG: a family of modular and scalable hierarchical ensemble methods to improve Gene Ontology term prediction.
Document Type
Article
Publication Date
12-7-2021
Original Citation
Notaro M,
Frasca M,
Petrini A,
Gliozzo J,
Casiraghi E,
Robinson P,
Valentini G.
HEMDAG: a family of modular and scalable hierarchical ensemble methods to improve Gene Ontology term prediction. Bioinformatics. 2021;37(23):4526-33
Keywords
JGM, Algorithms, Gene Ontology, Computational Biology, Proteins
JAX Source
Bioinformatics. 2021;37(23):4526-33
ISSN
1367-4811
PMID
34240108
DOI
https://doi.org/10.1093/bioinformatics/btab485
Grant
This work was supported by the Transition grant ‘UNIMI Partneriat H2020’ [PSR2015-1720GVALE_01] and by the PSR 2019 project ‘Machine Learning and Big Data Analysis for Bioinformatics’ [PSR2019_DIP_010_GVALE] funded by the University of Milano.
Abstract
MOTIVATION: Automated protein function prediction is a complex multi-class, multi-label, structured classification problem in which protein functions are organized in a controlled vocabulary, according to the Gene Ontology (GO). 'Hierarchy-unaware' classifiers, also known as 'flat' methods, predict GO terms without exploiting the inherent structure of the ontology, potentially violating the True-Path-Rule (TPR) that governs the GO, while 'hierarchy-aware' approaches, even if they obey the TPR, do not always show clear improvements with respect to flat methods, or do not scale well when applied to the full GO.
RESULTS: To overcome these limitations, we propose Hierarchical Ensemble Methods for Directed Acyclic Graphs (HEMDAG), a family of highly modular hierarchical ensembles of classifiers, able to build upon any flat method and to provide 'TPR-safe' predictions, by leveraging a combination of isotonic regression and TPR learning strategies. Extensive experiments on synthetic and real data across several organisms firstly show that HEMDAG can be used as a general tool to improve the predictions of flat classifiers, and secondly that HEMDAG is competitive versus state-of-the-art hierarchy-aware learning methods proposed in the last CAFA international challenges.
AVAILABILITY AND IMPLEMENTATION: Fully tested R code freely available at https://anaconda.org/bioconda/r-hemdag. Tutorial and documentation at https://hemdag.readthedocs.io.
SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.