Protein Function


Where we are actively engaged in the exciting field of function prediction for proteins. Our research in this area focuses on the development and implementation of advanced methodologies to predict Gene Ontology (GO) terms from protein sequences.

One of our notable contributions is the INGA method, which leverages multiple sources of information to generate consensus predictions for GO terms. INGA incorporates homology information, domain architecture, interaction networks, and valuable insights from the “dark proteome,” including transmembrane regions and intrinsically disordered regions. By integrating these diverse data sources, INGA achieves highly accurate predictions of protein functions.

In recognition of its performance, INGA was ranked among the top ten methods in the CAFA2 challenge and achieved an impressive second place in the CAFA3 blind test. The algorithm is capable of processing entire genomes in just a few hours, and its efficiency can be further enhanced by providing additional input files.

More recently, our research efforts have been focused on the analysis and prediction of functions associated with intrinsically disordered regions (IDRs). By harnessing cutting-edge machine learning technology, we have developed novel approaches that utilize protein embeddings and graph neural networks. These advancements enable us to unravel the functional implications of these dynamic and flexible regions, shedding light on the roles they play in cellular processes.

In addition to our research endeavors, we are actively involved in the Critical Assessment of Function Annotation (CAFA) initiative. As co-organizers, we contribute to the evaluation of function prediction methods and provide the evaluation software, CAFA-evaluator, used in the CAFA5 Challenge, which was hosted at Kaggle. This involvement allows us to stay at the forefront of the field, collaborate with leading researchers, and continuously refine our methodologies.

We are excited to share our expertise in function prediction and offer our advanced tools and methodologies to the scientific community. Whether you are interested in predicting protein functions from sequences, exploring the functional implications of intrinsically disordered regions, or participating in collaborative research initiatives, our lab is here to support your scientific endeavors.