Improvement of protein disorder prediction by brainstorming consensus

Deciphering the structure of a protein is essential towards understanding its functions. But there are proteins which undergo arbitrary changes in their tertiary structure. It is possible for a protein to be completely unstructured or be partially unstructured. Even though such proteins, known as intrinsically disordered proteins (IDP), do not have a fixed structure, they have an important biological role to play. We have designed a system which combines the results obtained from other state-of-the-art predictors to arrive at a single prediction. Most of the predictors have their own advantages and disadvantages. In this consensus scheme we have tried to overcome the disadvantages of one predictor by combining it with another. By combining 10 state-of-the art predictors we could achieve an increase of 15% in terms of recall on a dataset of 261 proteins.