fgga-package | R Documentation |
FGGA is a graph-based machine learning approach for the automated and consistent GO, PO, HPO and ZFA annotation of protein coding genes. The input is a set of ontoligical-terms annotated protein coding genes previously characterized in terms of a fixed number of user-defined features, including the presence/absence of PFAM domains, physical-chemical properties, presence of signal peptides, among others. The set of ontoligical terms defines the output cross-ontology subgraph. A hierarchical ensemble (SVMs) machine learning model is generated. This model can be used to predict the cross-ontology subgraph annotations of uncharacterized protein coding genes. Individual ontoligical-term annotations are accompanied by maximum a posteriori probability estimates issued by the native message passing algorithm of factor graphs.
Flavio E. Spetale, Paolo Cacchiarrelli and Elizabeth Tapia
BioInformatics
Cifasis-Conicet
spetale@cifasis-conicet.gov.ar
Maintainer: Flavio E. Spetale
Spetale F.E., et al. A Factor Graph Approach to Automated GO Annotation. PLoS ONE (2016). https://doi.org/10.1371/journal.pone.0146986.
Spetale Flavio E., et al. Consistent prediction of GO protein localization. Scientific Report (2018). https://doi.org/10.1038/s41598-018-26041-z.
fgga
, fgga2bipartite
, sumProduct
, svm
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.