What is it about?
We used high-performance computing to build a variety of heterogeneous ensembles to enhance the performance of protein function predictions. Heterogeneous ensembles tested include 8 Stacking models, one Ensemble Selection model (CES) and an unsupervised Mean aggregation model. Results show that several tested models indeed increase the performance in predicting 277 Gene Ontology terms.
Featured Image
Why is it important?
We enhanced the performance of functional predictions by combining diverse base predictors.
Perspectives
This is my first experience publishing my research work. It's a very meaningful progress to my professional career. This methodology is not only limited to protein function prediction, but it can also be applied to any kind of predictions. We also used it to build predictors for breast cancer, human phenotype ontology and so on.
Linhua Wang
Icahn School of Medicine at Mount Sinai
Read the Original
This page is a summary of: Large-scale protein function prediction using heterogeneous ensembles, F1000Research, September 2018, Faculty of 1000, Ltd.,
DOI: 10.12688/f1000research.16415.1.
You can read the full text:
Resources
Contributors
The following have contributed to this page
