scholarly journals Spatiotemporal identification of druggable binding sites using deep learning

2020 ◽  
Author(s):  
Igor Kozlovskii ◽  
Petr Popov
Keyword(s):  
Protein Binding ◽  
Binding Site ◽  
Binding Sites ◽  
Large Scale ◽  
Specific Binding ◽  
Three Dimensional ◽  
Growth Factor Receptor ◽  
Protein Binding Sites ◽  
Binding Site Identification ◽  
Scale Detection

Identification of novel protein binding sites expands «druggable genome» and opens new opportunities for drug discovery. Generally, presence or absence of a binding site depends on the three-dimensional conformation of a protein, making binding site identification resemble to object detection problem in computer vision. Here we introduce a computational approach for the large-scale detection of protein binding sites, named BiteNet, that considers protein conformations as the 3D-images, binding sites as the objects on these images to detect, and conformational ensembles of proteins as the 3D-videos to analyze. BiteNet is suitable for spatiotemporal detection of hard-to-spot allosteric binding sites, as we showed for conformation-specific binding site of the epidermal growth factor receptor, oligomer-specific binding site of the ion channel, and binding sites in G protein-coupled receptors. BiteNet outperforms state-of-the-art methods both in terms of accuracy and speed, taking about 1.5 minute to analyze 1000 conformations of a protein with 2000 atoms. BiteNet is available at https://github.com/i-Molecule/bitenet.

scholarly journals Spatiotemporal identification of druggable binding sites using deep learning

2020 ◽  
Vol 3 (1) ◽  
Author(s):  
Igor Kozlovskii ◽  
Petr Popov
Keyword(s):  
Protein Binding ◽  
Binding Site ◽  
Binding Sites ◽  
Large Scale ◽  
Specific Binding ◽  
Three Dimensional ◽  
Growth Factor Receptor ◽  
Protein Binding Sites ◽  
Binding Site Identification ◽  
Scale Detection

Abstract Identification of novel protein binding sites expands druggable genome and opens new opportunities for drug discovery. Generally, presence or absence of a binding site depends on the three-dimensional conformation of a protein, making binding site identification resemble the object detection problem in computer vision. Here we introduce a computational approach for the large-scale detection of protein binding sites, that considers protein conformations as 3D-images, binding sites as objects on these images to detect, and conformational ensembles of proteins as 3D-videos to analyze. BiteNet is suitable for spatiotemporal detection of hard-to-spot allosteric binding sites, as we showed for conformation-specific binding site of the epidermal growth factor receptor, oligomer-specific binding site of the ion channel, and binding site in G protein-coupled receptor. BiteNet outperforms state-of-the-art methods both in terms of accuracy and speed, taking about 1.5 minutes to analyze 1000 conformations of a protein with ~2000 atoms.

scholarly journals DeeplyTough: Learning Structural Comparison of Protein Binding Sites

2019 ◽  
Author(s):  
Martin Simonovsky ◽  
Joshua Meyers
Keyword(s):  
Protein Binding ◽  
Binding Sites ◽  
Protein Function ◽  
Large Scale ◽  
Protein Structures ◽  
Three Dimensional ◽  
Protein Binding Sites ◽  
Data Driven Approach ◽  
Benchmark Datasets ◽  
New Perspective

AbstractMotivationProtein binding site comparison (pocket matching) is of importance in drug discovery. Identification of similar binding sites can help guide efforts for hit finding, understanding polypharmacology and characterization of protein function. The design of pocket matching methods has traditionally involved much intuition, and has employed a broad variety of algorithms and representations of the input protein structures. We regard the high heterogeneity of past work and the recent availability of large-scale benchmarks as an indicator that a data-driven approach may provide a new perspective.ResultsWe propose DeeplyTough, a convolutional neural network that encodes a three-dimensional representation of protein binding sites into descriptor vectors that may be compared efficiently in an alignment-free manner by computing pairwise Euclidean distances. The network is trained with supervision: (i) to provide similar pockets with similar descriptors, (ii) to separate the descriptors of dissimilar pockets by a minimum margin, and (iii) to achieve robustness to nuisance variations. We evaluate our method using three large-scale benchmark datasets, on which it demonstrates excellent performance for held-out data coming from the training distribution and competitive performance when the trained network is required to generalize to datasets constructed independently.Availabilityhttps://github.com/BenevolentAI/[email protected],[email protected]

SitePrint: Three-Dimensional Pharmacophore Descriptors Derived from Protein Binding Sites for Family Based Active Site Analysis, Classification, and Drug Design.

ChemInform ◽  
2005 ◽  
Vol 36 (9) ◽  
Author(s):  
James R. Arnold ◽  
Keith W. Burdick ◽  
Scott C.-H. Pegg ◽  
Samuel Toba ◽  
Michelle L. Lamb ◽  
...  
Keyword(s):  
Drug Design ◽  
Protein Binding ◽  
Active Site ◽  
Binding Sites ◽  
Three Dimensional ◽  
Protein Binding Sites ◽  
Site Analysis ◽  
Family Based

scholarly journals Computational model predicts protein binding sites of a luminescent ligand equipped with guanidiniocarbonyl-pyrrole groups

2021 ◽  
Author(s):  
Neda Rafieiolhosseini ◽  
Matthias Killa ◽  
Niklas Tötsch ◽  
Jean-Noël Grad ◽  
Alexander Höing ◽  
...  
Keyword(s):  
Protein Binding ◽  
Computational Model ◽  
Binding Site ◽  
Binding Sites ◽  
Binding Proteins ◽  
Computational Approach ◽  
Hybrid Compound ◽  
Protein Binding Sites ◽  
Regulatory Processes ◽  
Enormous Number

The 14-3-3 protein family, one of the first discovered phosphoserine/phosphothreonine binding proteins, has attracted interest not only because of its important role in the cell regulatory processes but also due to its enormous number of interactions with other proteins. Here, we use a computational approach to find the binding sites of the designed hybrid compound featuring aggregation-induced emission luminophores as a potential supramolecular ligand for 14-3-3z in the presence and absence of C-Raf peptides. Our results suggest that the area above and below the central pore of the dimeric 14-3-3z protein is the most probable binding site for the ligand. Moreover, we predict that the position of the ligand is sensitive to the presence of phosphorylated C-Raf peptides.

CavSimBase: A Database for Large Scale Comparison of Protein Binding Sites

2016 ◽  
Vol 28 (6) ◽  
pp. 1423-1434 ◽  
Author(s):  
Matthias Leinweber ◽  
Thomas Fober ◽  
Marc Strickert ◽  
Lars Baumgartner ◽  
Gerhard Klebe ◽  
...  
Keyword(s):  
Protein Binding ◽  
Binding Sites ◽  
Large Scale ◽  
Protein Binding Sites

TORSIONAL RESONANCE MODE ATOMIC FORCE MICROSCOPY OF A PROTEIN–DNA COMPLEX

NANO ◽  
2008 ◽  
Vol 03 (06) ◽  
pp. 443-448 ◽  
Author(s):  
GILBERTO WEISSMÜLLER ◽  
AYHAN YURTSEVER ◽  
LILIAN T. COSTA ◽  
ANA B. F. PACHECO ◽  
PAULO M. BISCH ◽  
...  
Keyword(s):  
Protein Binding ◽  
Binding Site ◽  
Binding Sites ◽  
Physical Map ◽  
Resonance Mode ◽  
Dna Molecules ◽  
Image Processing Algorithm ◽  
Protein Binding Sites ◽  
Curved Dna ◽  
Atomic Force

Precise mapping of protein-binding sites on DNA is an important application of atomic force microscope (AFM) imaging. For a reliable measurement of distances on curved DNA molecules, an image-processing algorithm is required, which extracts the DNA contour from topographic AFM data. To this end we implemented an image analysis method providing an efficient way to obtain the contour together with a physical map of single and multiple protein-binding sites. This method relies on a calculation of the height profile along the DNA fragment, allowing one to determine the DNA length and the relative position of the binding site occupied by a protein. As a first test, complexes of the LexA repressor protein from the Escherichia coli SOS system and DNA fragments containing a specific LexA binding site (recA operator) were imaged by the torsional resonance mode (TR mode) and analyzed using the specialized algorithm. A topographic height of less than 0.5 nm of the DNA molecules indicates repulsive imaging conditions.

Nuclear protein-binding sites in a transcriptional control region of the rabbit alpha-globin gene

1993 ◽  
Vol 13 (9) ◽  
pp. 5439-5449
Author(s):  
S E Yost ◽  
B Shewchuk ◽  
R Hardison
Keyword(s):  
Protein Binding ◽  
Binding Site ◽  
Binding Sites ◽  
Nuclear Protein ◽  
Cpg Island ◽  
Globin Gene ◽  
Protein Binding Sites ◽  
Recognition Sites ◽  
Nuclear Extracts ◽  
Alpha Globin

The 5'-flanking and internal regions of the rabbit alpha-globin gene, which constitute a CpG island, are required for enhancer-independent expression in transfected cells. In this study, electrophoretic mobility shift assays revealed that a battery of nuclear proteins from both erythroid and nonerythroid cells bind specifically to these regulatory regions. Assays based on exonuclease III digestion, methylation interference, and DNase I footprinting identified sequences bound by proteins in crude nuclear extracts and by purified transcription factor Sp1. In the 5' flank, recognition sites for the transcription factors alpha-IRP (positions -53 to -44 relative to the cap site), CP1 (-73 to -69), and Sp1 (-95 to -90) are bound by proteins in K562 cell nuclear extracts, as are three extended upstream regions. Two recognition sites for Sp1 in intron 1 are also bound both by proteins in crude nuclear extracts and by purified Sp1. The sequences CCAC in intron 2 and C5 in the 3'-untranslated region also bind proteins. A major binding site found in exon 1, TATGGCGC, matches in sequence and methylation interference pattern the binding site for nuclear protein YY1, and binding is inhibited through competition by YY1-specific oligonucleotides. The protein-binding sites flanking and internal to the rabbit alpha-globin gene may form an extended promoter.

scholarly journals Important Roles for E Protein Binding Sites within the Immunoglobulin κ Chain Intronic Enhancer in Activating Vκ Jκ Rearrangement

2004 ◽  
Vol 200 (9) ◽  
pp. 1205-1211 ◽  
Author(s):  
Matthew A. Inlay ◽  
Hua Tian ◽  
Tongxiang Lin ◽  
Yang Xu
Keyword(s):  
Homologous Recombination ◽  
Protein Binding ◽  
Binding Site ◽  
Light Chain ◽  
Binding Sites ◽  
Apparent Effect ◽  
Protein Binding Sites ◽  
Intronic Enhancer ◽  
E Protein ◽  
E Boxes

The immunoglobulin κ light chain intronic enhancer (iEκ) activates κ rearrangement and is required to maintain the earlier or more efficient rearrangement of κ versus lambda (λ). To understand the mechanism of how iEκ regulates κ rearrangement, we employed homologous recombination to mutate individual functional motifs within iEκ in the endogenous κ locus, including the NF-κB binding site (κB), as well as κE1, κE2, and κE3 E boxes. Analysis of the impacts of these mutations revealed that κE2 and to a lesser extent κE1, but not κE3, were important for activating κ rearrangement. Surprisingly, mutation of the κB site had no apparent effect on κ rearrangement. Comparable to the deletion of the entire iEκ, simultaneous mutation of κE1 and κE2 reduces the efficiency of κ rearrangement much more dramatically than either κE1 or κE2 mutation alone. Because E2A family proteins are the only known factors that bind to these E boxes, these findings provide unambiguous evidence that E2A is a key regulator of κ rearrangement.

Sign in / Sign up

Export Citation Format

Share Document