About this Research Topic
This Research Topic covers the recent development of deep learning methods for predicting and determining protein structures and interactions.
Protein function is largely determined by protein structures and interactions. Elucidating the structures and interactions of proteins has fundamental impacts on many biomedical domains. Because only a tiny fraction of protein structures and interactions can be determined by experimental techniques such as x-ray crystallography, NMR and cryo-EM to due to the high cost of experimental structure determination, the development of high-throughput computational methods for prediction of protein structures and interactions has become an important research topic in bioinformatics, computational biology and structural biology. Computational modeling methods have the great potentials to not only predict structures and interactions from protein sequences at very low cost and high speed, but also to improve the efficiency and effectiveness of protein structure determination.
The recent major advances in the field of computational prediction and determination of protein structures and interactions are largely driven by the application of advanced of deep learning methods to extracting structural patterns from the large amount of protein sequence and structural data and to analyze experimental structural data such as cryo-EM images. For instance, the recent revolutionary advances that drastically improve the accuracy and coverage of ab initio protein structure prediction as demonstrated in the 13th Critical Assessment of Techniques for Protein Structure Prediction (CASP13) are due to the development of deep learning methods of predicting residue-residue contacts and distances leveraging residue-residue co-evolutionary signals in protein sequences.
In order to present the recent exciting advances in the technological development of deep learning methods for predicting and determining protein structures and interactions in a centralized and easily accessible venue, we are pleased to organize a special issue with Frontiers in Bioinformatics on this important topic. The articles on this topic may touch various aspects of prediction and determination of protein structures and interactions ranging from prediction of 1D and 2D structural features, tertiary structures, and quaternary structures and interactions, assessment of the quality of structural models, and analysis of experimental structural data. Particularly, we welcome the submission of deep learning methods of addressing the following problems:
• Prediction of protein structural features (e.g., secondary structures, solvent accessibilities, torsion angles, disorders, domain boundaries, and binding and interaction sites)
• Prediction of intra-chain or inter-chain residue-residue contacts, distances, hydrogen bonds, disulfide bonds, and residue-residue orientations
• Prediction of protein tertiary structures
• Prediction of protein interactions or quaternary structures
• Assessment of the quality of protein tertiary or quaternary structural models
• Analysis of experimental structural data (e.g., determining / predicting secondary structures and tertiary structures from cryo-EM image data, experimental data assisted protein structure modeling)
Different article types including Original Research, Reviews, Perspective, and Methods will be considered.
Protein function is largely determined by protein structures and interactions. Elucidating the structures and interactions of proteins has fundamental impacts on many biomedical domains. Because only a tiny fraction of protein structures and interactions can be determined by experimental techniques such as x-ray crystallography, NMR and cryo-EM to due to the high cost of experimental structure determination, the development of high-throughput computational methods for prediction of protein structures and interactions has become an important research topic in bioinformatics, computational biology and structural biology. Computational modeling methods have the great potentials to not only predict structures and interactions from protein sequences at very low cost and high speed, but also to improve the efficiency and effectiveness of protein structure determination.
The recent major advances in the field of computational prediction and determination of protein structures and interactions are largely driven by the application of advanced of deep learning methods to extracting structural patterns from the large amount of protein sequence and structural data and to analyze experimental structural data such as cryo-EM images. For instance, the recent revolutionary advances that drastically improve the accuracy and coverage of ab initio protein structure prediction as demonstrated in the 13th Critical Assessment of Techniques for Protein Structure Prediction (CASP13) are due to the development of deep learning methods of predicting residue-residue contacts and distances leveraging residue-residue co-evolutionary signals in protein sequences.
In order to present the recent exciting advances in the technological development of deep learning methods for predicting and determining protein structures and interactions in a centralized and easily accessible venue, we are pleased to organize a special issue with Frontiers in Bioinformatics on this important topic. The articles on this topic may touch various aspects of prediction and determination of protein structures and interactions ranging from prediction of 1D and 2D structural features, tertiary structures, and quaternary structures and interactions, assessment of the quality of structural models, and analysis of experimental structural data. Particularly, we welcome the submission of deep learning methods of addressing the following problems:
• Prediction of protein structural features (e.g., secondary structures, solvent accessibilities, torsion angles, disorders, domain boundaries, and binding and interaction sites)
• Prediction of intra-chain or inter-chain residue-residue contacts, distances, hydrogen bonds, disulfide bonds, and residue-residue orientations
• Prediction of protein tertiary structures
• Prediction of protein interactions or quaternary structures
• Assessment of the quality of protein tertiary or quaternary structural models
• Analysis of experimental structural data (e.g., determining / predicting secondary structures and tertiary structures from cryo-EM image data, experimental data assisted protein structure modeling)
Different article types including Original Research, Reviews, Perspective, and Methods will be considered.
Keywords: Deep learning, protein structure prediction, protein interaction prediction, protein structure determination
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
