Large-scale automated protein function prediction and an experimental validation

What is it about?

Our study is an improved automated protein function prediction method (UniGOPred) based on Gene Ontology (GO) terms. Users can submit their protein sequences as input to get predictions for the query protein. We also provide pre-computed protein function prediction for UniProtKB protein sequences and sequences of model organisms. We also performed experimental validation of our predictions on PTEN and CHD8 proteins.

Featured Image

Why is it important?

First, It is not possible to use most of the protein function prediction methods as they do not provide their methods as an webserver or open-access tool. We provided UniGOPred, as an open-access webserver to be used by scientific community. Second, our study provide not only positive predictions but also negative predictions of Gene Ontology Terms. In addition, especially our negative training dataset creation method helps to create more reliable negative training dataset. Experimental validation of the prediction is quite scarce in the literature. Here, we also provided experimental validation of our predictions for PTEN and CHD8 proteins and their transcript variants.

Perspectives