New machine learning methods for prediction of protein secondary structures

The first and probably the most important step in predicting the tertiary structure of proteins from its primary structure is to predict as many as possible secondary structures in a protein chain. Secondary structure prediction problem has been known for almost a quarter of century. In this paper, new machine learning methods such as LAD, LEM2, and MODLEM have been used for secondary protein structure prediction with the aim to choose the best among them which will be later parallelized in order to handle a huge amount of data sets.