scholarly journals Improved antibody structure prediction by deep learning of side chain conformations

2021 ◽  
Author(s):  
Deniz Akpinaroglu ◽  
Jeffrey A Ruffolo ◽  
Sai Pooja Mahajan ◽  
Jeffrey J. Gray
Keyword(s):  
Deep Learning ◽  
Structure Prediction ◽  
Antibody Engineering ◽  
Side Chain ◽  
Antibody Structure ◽  
Backbone Structure ◽  
Complementarity Determining Regions ◽  
Chain Conformations ◽  
Antibody Function ◽  
Antigenic Epitopes

Antibody engineering is becoming increasingly popular in the medical field for the development of diagnostics and immunotherapies. Antibody function relies largely on the recognition and binding of antigenic epitopes via the loops in the complementarity determining regions. Hence, accurate high-resolution modeling of these loops is essential for effective antibody engineering and design. Deep learning methods have previously been shown to effectively predict antibody backbone structures described as a set of inter-residue distances and orientations. However, antigen binding is also dependent on the specific conformations of surface side chains. To address this shortcoming, we created DeepSCAb: a deep learning method that predicts inter-residue geometries as well as side chain dihedrals of the antibody variable fragment. The network requires only sequence as input, rendering our method particularly useful for antibodies without any known backbone conformations. Rotamer predictions use an interpretable self-attention layer, which learns to identify structurally conserved anchor positions across several species. We evaluate the performance of our model for discriminating near-native structures from sets of decoys and find that DeepSCAb outperforms similar methods lacking side chain context. When compared to alternative rotamer repacking methods, which require an input backbone structure, DeepSCAb predicts side chain conformations competitively. Our findings suggest that DeepSCAb improves antibody structure prediction with accurate side chain modeling and is adaptable to applications in docking of antibody-antigen complexes and design of new therapeutic antibody sequences.

scholarly journals DLPacker: Deep Learning for Prediction of Amino Acid Side Chain Conformations in Proteins

2021 ◽  
Author(s):  
Mikita Misiura ◽  
Raghav Shroff ◽  
Ross Thyer ◽  
Anatoly Kolomeisky
Keyword(s):  
Amino Acids ◽  
Amino Acid ◽  
Protein Design ◽  
Structure Prediction ◽  
Deep Neural Networks ◽  
Side Chain ◽  
Amino Acid Side Chain ◽  
Image Transformation ◽  
Hydrophobic Amino Acids ◽  
Chain Conformations

Prediction of side chain conformations of amino acids in proteins (also termed 'packing') is an important and challenging part of protein structure prediction with many interesting applications in protein design. A variety of methods for packing have been developed but more accurate ones are still needed. Machine learning (ML) methods have recently become a powerful tool for solving various problems in diverse areas of science, including structural biology. In this work we evaluate the potential of Deep Neural Networks (DNNs) for prediction of amino acid side chain conformations. We formulate the problem as image-to-image transformation and train a U-net style DNN to solve the problem. We show that our method outperforms other physics-based methods by a significant margin: reconstruction RMSDs for most amino acids are about 20% smaller compared to SCWRL4 and Rosetta Packer with RMSDs for bulky hydrophobic amino acids Phe, Tyr and Trp being up to 50% smaller.

scholarly journals ABlooper: Fast accurate antibody CDR loop structure prediction with accuracy estimation

2021 ◽  
Author(s):  
Brennan Abanades ◽  
Guy Georges ◽  
Alexander Bujotzek ◽  
Charlotte M Deane
Keyword(s):  
Deep Learning ◽  
Structure Prediction ◽  
Structural Variation ◽  
Protein Structures ◽  
Step Change ◽  
Antigen Binding ◽  
Loop Structure ◽  
Accurate Knowledge ◽  
Complementarity Determining Regions ◽  
Confidence Estimate

Antibodies are a key component of the immune system and have been extensively used as biotherapeutics. Accurate knowledge of their structure is central to understanding their function. The key area for antigen binding and the main area of structural variation in antibodies are concentrated in their six complementarity determining regions (CDRs), with the most variable being the CDR-H3 loop. The sequence and structure variability of CDR-H3 make it particularly challenging to model. Recently, deep learning methods have offered a step change in our ability to predict protein structures. In this work we present ABlooper, an end-toend equivariant deep-learning based CDR loop structure prediction tool. ABlooper predicts the structure of CDR loops with high accuracy and provides a confidence estimate for each of its predictions. On the models of the Rosetta Antibody Benchmark, ABlooper makes predictions with an average H3 RMSD of 2.45Å, which drops to 2.02Å when considering only its 76% most confident predictions.

scholarly journals Antibody structure prediction using interpretable deep learning

Patterns ◽  
2021 ◽  
pp. 100406
Author(s):  
Jeffrey A. Ruffolo ◽  
Jeremias Sulam ◽  
Jeffrey J. Gray
Keyword(s):  
Deep Learning ◽  
Structure Prediction ◽  
Antibody Structure

scholarly journals How Communication Pathways Bridge Local and Global Conformations in an IgG4 Antibody: a Molecular Dynamics Study

2021 ◽  
Author(s):  
Thomas Tarenzi ◽  
Marta Rigoli ◽  
Raffaello Potestio
Keyword(s):  
Molecular Dynamics ◽  
Conformational Changes ◽  
Md Simulations ◽  
Specific Sequence ◽  
Multi Scale ◽  
Antibody Structure ◽  
Complementarity Determining Regions ◽  
Global And Local ◽  
Cell Death Protein ◽  
Antibody Function

The affinity of an antibody for its antigen is primarily determined by the specific sequence and structural arrangement of the complementarity-determining regions (CDRs). Recently, however, evidence has accumulated that points toward a nontrivial relation between the CDR and distal sites on the antibody structure: variations in the binding strengths have been observed upon mutating amino acids separated from the paratope by several nanometers, thus suggesting the existence of a communication network within antibodies whose extension and relevance might be deeper than insofar expected. In this work, we test this hypothesis by means of molecular dynamics (MD) simulations of the IgG4 monoclonal antibody pembrolizumab, an approved drug that targets the programmed cell death protein 1 (PD-1). The molecule is simulated in both the apo and holo states, totalling 4 μs of MD trajectory. The analysis of these simulations shows that the bound antibody explores a restricted range of conformations with respect to the apo one, and that the global conformation of the molecule correlates with that of the CDR; a pivotal role in this relationship is played by the relatively short hinge, which mechanically couples Fab and Fc domains. These results support the hypothesis that pembrolizumab behaves as a complex machinery, with a multi-scale hierarchy of global and local conformational changes that communicate with one another. The analysis pipeline developed in this work is general, and it can help shed further light on the mechanistic aspects of antibody function.

Crystal Structure Prediction via Deep Learning

2018 ◽  
Vol 140 (32) ◽  
pp. 10158-10168 ◽  
Author(s):  
Kevin Ryan ◽  
Jeff Lengyel ◽  
Michael Shatruk
Keyword(s):  
Crystal Structure ◽  
Deep Learning ◽  
Structure Prediction ◽  
Crystal Structure Prediction

scholarly journals Template-based prediction of protein structure with deep learning

BMC Genomics ◽  
2020 ◽  
Vol 21 (S11) ◽  
Author(s):  
Haicang Zhang ◽  
Yufeng Shen
Keyword(s):  
Deep Learning ◽  
Protein Structure ◽  
Structure Prediction ◽  
Tertiary Structure ◽  
Query Sequence ◽  
Dynamic Programming Algorithm ◽  
Tertiary Structure Prediction ◽  
Protein Tertiary Structure ◽  
Protein Threading ◽  
Protein Tertiary Structure Prediction

Abstract Background Accurate prediction of protein structure is fundamentally important to understand biological function of proteins. Template-based modeling, including protein threading and homology modeling, is a popular method for protein tertiary structure prediction. However, accurate template-query alignment and template selection are still very challenging, especially for the proteins with only distant homologs available. Results We propose a new template-based modelling method called ThreaderAI to improve protein tertiary structure prediction. ThreaderAI formulates the task of aligning query sequence with template as the classical pixel classification problem in computer vision and naturally applies deep residual neural network in prediction. ThreaderAI first employs deep learning to predict residue-residue aligning probability matrix by integrating sequence profile, predicted sequential structural features, and predicted residue-residue contacts, and then builds template-query alignment by applying a dynamic programming algorithm on the probability matrix. We evaluated our methods both in generating accurate template-query alignment and protein threading. Experimental results show that ThreaderAI outperforms currently popular template-based modelling methods HHpred, CNFpred, and the latest contact-assisted method CEthreader, especially on the proteins that do not have close homologs with known structures. In particular, in terms of alignment accuracy measured with TM-score, ThreaderAI outperforms HHpred, CNFpred, and CEthreader by 56, 13, and 11%, respectively, on template-query pairs at the similarity of fold level from SCOPe data. And on CASP13’s TBM-hard data, ThreaderAI outperforms HHpred, CNFpred, and CEthreader by 16, 9 and 8% in terms of TM-score, respectively. Conclusions These results demonstrate that with the help of deep learning, ThreaderAI can significantly improve the accuracy of template-based structure prediction, especially for distant-homology proteins.

Stereospecific Association of C-20 Epimers of 3β-Hydroxy-16-oxo-24-nor-17-azachol-5-eno-23-nitryle

1997 ◽  
Vol 52 (6) ◽  
pp. 749-756
Author(s):  
Zofia Plesnar ◽  
Stanisław Malanowski ◽  
Zenon Lotowski ◽  
Jacek W. Morzycki ◽  
Jadwiga Frelek ◽  
...  
Keyword(s):  
Proton Nuclear Magnetic Resonance ◽  
Side Chain ◽  
Water Molecules ◽  
Carbonyl Groups ◽  
Nmr Data ◽  
Molecular Modelling Studies ◽  
Modelling Studies ◽  
Lactam Carbonyl ◽  
Self Association ◽  
Chain Conformations

Abstract The cryoscopic measurements show that title compounds are strongly associated in CHCl3 solutions. The association of the 20 R epimer is distinctly less pronounced than that of the 20 S epipmer. Self-association of the 20 S epimer leads to the formation of very large com­plexes. The 20 R epimer forms associates via water molecules. The dissimilarity may be ex­plained in terms of different accessibility of the lactam carbonyl groups in the two epimers for the association. It is proposed that the association process is controlled by the configura­tion at the carbon atom C(20) and conformation around the C(20)-C(22) bond. Populations of side chain conformations of both epimers were determined by means of proton nuclear magnetic resonance. It was found for the 20 R epimer that the t and the -g rotamers are almost equally populated, and the rotamer +g is excluded. For the 20 S epimer the +g rotamer predominates over the t one, and the -g rotamer is excluded. The NMR data are fully consistent with the results of the molecular modelling studies.

Sign in / Sign up

Export Citation Format

Share Document