Ensemble of Protein Stability upon Point Mutation Predictors

• Eduardo Kenji Hasegawa de Freitas FURG https://orcid.org/0000-0002-1476-4779
• Alex Dias Camargo FURG https://orcid.org/0000-0002-6377-9379
• Maurício Balboni FURG https://orcid.org/0000-0001-7653-0962
• Adriano V. Werhli FURG https://orcid.org/0000-0001-7107-5024
• Karina dos Santos Machado FURG https://orcid.org/0000-0002-8966-5708

Resumo

Computational determination of protein stability upon point mutations is very useful in a wide field of applications. The reliability of such computational predictions is essential. Unfortunately, existing computational tools frequently disagree in their results. In the present study, the usage of Ensemble Learning Algorithms to aggregate the results from different stability prediction tools is investigated. Techniques of Stacking, Bagging, and Boosting as well as different Machine Learning algorithms as combiner function are explored. All the investigation is carried out in real dataset ProTherm for which experimental results are known. The proposed methodology was validated considering two different experiments according to the training set. Results show that our proposed ensemble approach is appropriate to predict the effect of point mutations on protein stability showing more reliable results than the individual tools improving overall accuracy, precision, and/or recall.

Palavras-chave: Ensemble learning, Bioinformatics, Protein stability, Point mutations