Rhys Heffernan

TL;DR: The accuracy of the iterative use of predicted secondary structure and backbone torsion angles and dihedrals based on Cα atoms is higher than those of model structures from current state-of-the-art techniques, suggesting the potentially beneficial use of these predicted properties for model assessment and ranking.

TL;DR: The application of LSTM‐BRNN to the prediction of protein structural properties makes the most significant improvement for residues with the most long‐range contacts over a previous window‐based, deep‐learning method SPIDER2.

TL;DR: This study proposes two segmentation based feature extraction methods and shows that by applying a Support Vector Machine (SVM) classifier to the extracted features, they are able to enhance Gram-positive and Gram-negative subcellular localization prediction accuracies by up to 6.4% better than previous studies including the studies that used GO for feature extraction.

TL;DR: The time has come to finish off the final stretch of the long march towards protein secondary structure prediction as more powerful deep learning methods with improved capability of capturing long-range interactions begin to emerge as the next generation of techniques forsecondary structure prediction.

TL;DR: A newly developed method SPIDER2 is described that utilizes three iterations of deep learning neural networks to improve the prediction accuracy of several structural properties simultaneously and provides state-of-the-art, all-in-one accurate prediction of local structure and solvent accessible surface area.