Combining evolutionary and structural information for local protein structure prediction

AbstractThe Protein Data Bank (PDB) undergoes an exponential expansion in terms of the number of macromolecular structures deposited every year. A pivotal question is how this rapid growth of structural information improves the quality of three-dimensional models constructed by contemporary bioinformatics approaches. To address this problem, we performed a retrospective analysis of the structural coverage of a representative set of proteins using remote homology detected by COMPASS and HHpred. We show that the number of proteins whose structures can be confidently predicted increased during a 9-year period between 2005 and 2014 on account of the PDB growth alone. Nevertheless, this encouraging trend slowed down noticeably around the year 2008 and has yielded insignificant improvements ever since. At the current pace, it is unlikely that the protein structure prediction problem will be solved in the near future using existing template-based modeling techniques. Therefore, further advances in experimental structure determination, qualitatively better approaches in fold recognition, and more accurate template-free structure prediction methods are desperately needed.

Download Full-text

Protein Structure Prediction Using a Maximum Likelihood Formulation of a Recurrent Geometric Network

10.1101/2021.09.03.458873 ◽

2021 ◽

Author(s):

Guowei Qi ◽

Mallory R Tollefson ◽

Rose A Gogal ◽

Richard J.H. Smith ◽

Mohammed AlQuraishi ◽

...

Keyword(s):

Protein Structure ◽

Maximum Likelihood ◽

Protein Structure Prediction ◽

Structure Prediction ◽

Structural Information ◽

Atomic Displacement ◽

Machine Learning Algorithms ◽

Human Proteome ◽

Training Dataset ◽

Model Training

Only ~40% of the human proteome has structural coordinates available from experiment (i.e., X-ray crystallography, NMR spectroscopy, or cryo-EM) or homology modeling with quality templates (i.e., 30% sequence identity or greater), leaving most of the proteome structurally unsolved. Deep learning (DL) methods for predicting protein structure can help close knowledge gaps where experimental and homology models are difficult to obtain. Recent advances in these DL methods have shown promising results in expanding structural coverage to the scale of the entire human proteome, providing researchers with more complete protein structural information. Here, we improve upon an existing DL algorithm for protein structure prediction, the Recurrent Geometric Network (RGN). We first expand the training dataset to include experimental uncertainty data in the form of atomic displacement parameters, then derive a maximum likelihood loss function that incorporates this uncertainty data into model training. Compared to the original RGN, our novel maximum likelihood model improves the rate of convergence of initial model training and ultimately results in more accurate structure prediction according to the root mean square deviation (RMSD) of backbone atoms, the Global Distance Test (GDT), the Global Distance Test High Accuracy (GDT-HA), and the Template-Modeling Score (TM-Score). Our model also predicts structures with more favorable backbone torsions, which provide more accurate starting coordinates for downstream physics-based simulations. Based on these results, our maximum likelihood reformulation provides a framework for improving existing or future machine learning algorithms for protein structure prediction. The augmented dataset, data collection scripts, reformulated RGN source code, and a series of trained models are publicly available at https://github.com/SchniedersLab/likelihood-rgn.

Download Full-text

Complementing sequence-derived features with structural information extracted from fragment libraries for protein structure prediction

BMC Bioinformatics ◽

10.1186/s12859-021-04258-6 ◽

2021 ◽

Vol 22 (1) ◽

Author(s):

Siyuan Liu ◽

Tong Wang ◽

Qijiang Xu ◽

Bin Shao ◽

Jian Yin ◽

...

Keyword(s):

Protein Folding ◽

Protein Structure ◽

Protein Structure Prediction ◽

Structure Prediction ◽

Structural Information ◽

Three Dimensional ◽

Structural Features ◽

Dimensional Structure ◽

Fragment Assembly ◽

Fragment Libraries

Abstract Background Fragment libraries play a key role in fragment-assembly based protein structure prediction, where protein fragments are assembled to form a complete three-dimensional structure. Rich and accurate structural information embedded in fragment libraries has not been systematically extracted and used beyond fragment assembly. Methods To better leverage the valuable structural information for protein structure prediction, we extracted seven types of structural information from fragment libraries. We broadened the usage of such structural information by transforming fragment libraries into protein-specific potentials for gradient-descent based protein folding and encoding fragment libraries as structural features for protein property prediction. Results Fragment libraires improved the accuracy of protein folding and outperformed state-of-the-art algorithms with respect to predicted properties, such as torsion angles and inter-residue distances. Conclusion Our work implies that the rich structural information extracted from fragment libraries can complement sequence-derived features to help protein structure prediction.

Download Full-text

Sequence Specific Dihedral Angle Distribution: Application in Protein Structure Prediction and Evaluation

Plant Tissue Culture and Biotechnology ◽

10.3329/ptcb.v19i2.5439 ◽

1970 ◽

Vol 19 (2) ◽

pp. 217-226

Author(s):

S. M. Minhaz Ud-Dean ◽

Mahdi Muhammad Moosa

Keyword(s):

Protein Structure ◽

Dihedral Angle ◽

Protein Structure Prediction ◽

Structure Prediction ◽

Protein Structures ◽

Angle Distribution ◽

Ramachandran Plot ◽

Specific Data ◽

Specific Distribution ◽

Structure Evaluation

Protein structure prediction and evaluation is one of the major fields of computational biology. Estimation of dihedral angle can provide information about the acceptability of both theoretically predicted and experimentally determined structures. Here we report on the sequence specific dihedral angle distribution of high resolution protein structures available in PDB and have developed Sasichandran, a tool for sequence specific dihedral angle prediction and structure evaluation. This tool will allow evaluation of a protein structure in pdb format from the sequence specific distribution of Ramachandran angles. Additionally, it will allow retrieval of the most probable Ramachandran angles for a given sequence along with the sequence specific data. Key words: Torsion angle, φ-ψ distribution, sequence specific ramachandran plot, Ramasekharan, protein structure appraisal D.O.I. 10.3329/ptcb.v19i2.5439 Plant Tissue Cult. & Biotech. 19(2): 217-226, 2009 (December)

Download Full-text