scholarly journals GNINA 1.0: molecular docking with deep learning

2021 ◽  
Vol 13 (1) ◽  
Author(s):  
Andrew T. McNutt ◽  
Paul Francoeur ◽  
Rishal Aggarwal ◽  
Tomohide Masuda ◽  
Rocco Meli ◽  
...  
Keyword(s):  
Molecular Docking ◽  
Root Mean Square ◽  
Root Mean Square Deviation ◽  
Computational Cost ◽  
Scoring Function ◽  
Binding Pocket ◽  
Protein Docking ◽  
Mean Square ◽  
Mean Square Deviation ◽  
Autodock Vina

AbstractMolecular docking computationally predicts the conformation of a small molecule when binding to a receptor. Scoring functions are a vital piece of any molecular docking pipeline as they determine the fitness of sampled poses. Here we describe and evaluate the 1.0 release of the Gnina docking software, which utilizes an ensemble of convolutional neural networks (CNNs) as a scoring function. We also explore an array of parameter values for Gnina 1.0 to optimize docking performance and computational cost. Docking performance, as evaluated by the percentage of targets where the top pose is better than 2Å root mean square deviation (Top1), is compared to AutoDock Vina scoring when utilizing explicitly defined binding pockets or whole protein docking. Gnina, utilizing a CNN scoring function to rescore the output poses, outperforms AutoDock Vina scoring on redocking and cross-docking tasks when the binding pocket is defined (Top1 increases from 58% to 73% and from 27% to 37%, respectively) and when the whole protein defines the binding pocket (Top1 increases from 31% to 38% and from 12% to 16%, respectively). The derived ensemble of CNNs generalizes to unseen proteins and ligands and produces scores that correlate well with the root mean square deviation to the known binding pose. We provide the 1.0 version of Gnina under an open source license for use as a molecular docking tool at https://github.com/gnina/gnina.

scholarly journals GNINA 1.0: Molecular Docking with Deep Learning

2021 ◽  
Author(s):  
Andrew McNutt ◽  
Paul Francoeur ◽  
Rishal Aggarwal ◽  
Tomohide Masuda ◽  
Rocco Meli ◽  
...  
Keyword(s):  
Molecular Docking ◽  
Root Mean Square ◽  
Root Mean Square Deviation ◽  
Computational Cost ◽  
Scoring Function ◽  
Binding Pocket ◽  
Protein Docking ◽  
Mean Square ◽  
Mean Square Deviation ◽  
Autodock Vina

Molecular docking computationally predicts the conformation of a small molecule when binding to a receptor. Scoring functions are a vital piece of any molecular docking pipeline as they determine the fitness of sampled poses. Here we describe and evaluate the 1.0 release of the Gnina docking software, which utilizes an ensemble of convolutional neural networks (CNNs) as a scoring function. We also explore an array of parameter values for Gnina 1.0 to optimize docking performance and computational cost. Docking performance, as evaluated by the percentage of targets where the top pose is better than 2A root mean square deviation (Top1), is compared to AutoDock Vina scoring when utilizing explicitly defined binding pockets or whole protein docking. Gnina, utilizing a CNN scoring function to rescore the output poses, outperforms AutoDock Vina scoring on redocking and cross-docking tasks when the binding pocket is defined (Top1 increases from 58% to 73% and from 27% to 37%, respectively) and when the whole protein defines the binding pocket (Top1 increases from 31% to 38% and from 12% to 16%, respectively). The derived ensemble of CNNs generalizes to unseen proteins and ligands and produces scores that correlate well with the root mean square deviation to the known binding pose. We provide the 1.0 version of Gnina under and open source license for use as a molecular docking tool at https://github.com/gnina/gnina.

scholarly journals GNINA 1.0: Molecular Docking with Deep Learning

2021 ◽  
Author(s):  
Andrew McNutt ◽  
Paul Francoeur ◽  
Rishal Aggarwal ◽  
Tomohide Masuda ◽  
Rocco Meli ◽  
...  
Keyword(s):  
Molecular Docking ◽  
Root Mean Square ◽  
Root Mean Square Deviation ◽  
Computational Cost ◽  
Scoring Function ◽  
Binding Pocket ◽  
Protein Docking ◽  
Mean Square ◽  
Mean Square Deviation ◽  
Autodock Vina

Molecular docking computationally predicts the conformation of a small molecule when binding to a receptor. Scoring functions are a vital piece of any molecular docking pipeline as they determine the fitness of sampled poses. Here we describe and evaluate the 1.0 release of the Gnina docking software, which utilizes an ensemble of convolutional neural networks (CNNs) as a scoring function. We also explore an array of parameter values for Gnina 1.0 to optimize docking performance and computational cost. Docking performance, as evaluated by the percentage of targets where the top pose is better than 2A root mean square deviation (Top1), is compared to AutoDock Vina scoring when utilizing explicitly defined binding pockets or whole protein docking. Gnina, utilizing a CNN scoring function to rescore the output poses, outperforms AutoDock Vina scoring on redocking and cross-docking tasks when the binding pocket is defined (Top1 increases from 58% to 73% and from 27% to 37%, respectively) and when the whole protein defines the binding pocket (Top1 increases from 31% to 38% and from 12% to 16%, respectively). The derived ensemble of CNNs generalizes to unseen proteins and ligands and produces scores that correlate well with the root mean square deviation to the known binding pose. We provide the 1.0 version of Gnina under and open source license for use as a molecular docking tool at https://github.com/gnina/gnina.

POTENSI DAUN ASAM JAWA (Tamarindus indica L.) SEBAGAI ALTERNATIF ANTIINFLAMASI: STUDI IN SILICO

2019 ◽  
pp. 42-50
Author(s):  
Erma Yunita ◽  
Siti Fatimah ◽  
Deni Yulianto ◽  
Vedy Trikuncahyo ◽  
Zihan Khodijah
Keyword(s):  
Molecular Docking ◽  
Root Mean Square ◽  
Root Mean Square Deviation ◽  
In Silico ◽  
Mean Square ◽  
Mean Square Deviation ◽  
Tamarindus Indica ◽  
Protein Target ◽  
Cox 2 ◽  
Cox 1

  Daun asam jawa (Tamarindus indica L.) merupakan tanaman yang memiliki banyak khasiat. Kandungan senyawa kimia yang terkandung salah satunya Kuersetin. Kuersetin merupakan senyawa flavonoid yang dapat digunakan sebagai anti inflamasi. Penelitian ini bertujuan untuk mengetahui potensi aktivitas Kuersetin dari daun asam jawa sebagai anti inflamasi terhadap protein COX-1 dan COX-2 secara in silico. Ekstrak daun asam jawa diperoleh dengan maserasi bertingkat menggunakan heksan dan etanol. Kadar Kuersetinnya dihitung secara spektrofotometri UVVis. Konfirmasi aktivitas antiinflamasi dilakukan secara in silico. Protein yang digunakan adalah 6COX, 3PGH, dan 1EQH. Kuersetin sebagai senyawa aktif sedangkan Aspirin digunakan sebagai zat pembanding. Preparasi ligan Kuersetin menggunakan MarvinSketch kemudian preparasi protein target 6COX, 1EQH, dan 3PGH menggunakan YASARA. Selanjutnya melakukan molecular docking menggunakan program PLANTS. Parameter evaluasi validasi dapat dilihat dari nilai Root Mean Square Deviation (RMSD), dimana nilai RMSD yang diterima adalah kurang dari 2Å. Kadar Kuersetin yang diperoleh dalam ekstrak dalam daun asam jawa sebesar 31,26 mg/g. Hasil docking menunjukkan bahwa Kuersetin mampu berinteraksi dengan 1EQH, 3PGH, dan 6COX dimana skor dockingnya masing-masing adalah -77,6195; -75,1344; dan -82,2454, sedangkan hasil docking Aspirin masing-masing adalah -69,8784; -75,2421; dan - 72,0884. Kuersetin memiliki potensi sebagai anti inflamasi yang lebih baik dibandingkan dengan Aspirin namun memiliki resiko lebih tinggi menyebabkan ulkus lambung dibanding Aspirin.

Fast collocation for Moho estimation from GOCE gravity data: the Iran case study

2020 ◽  
Vol 221 (1) ◽  
pp. 651-664
Author(s):  
H Heydarizadeh Shali ◽  
D Sampietro ◽  
A Safari ◽  
M Capponi ◽  
A Bahroudi
Keyword(s):  
Root Mean Square ◽  
Root Mean Square Deviation ◽  
Seismic Data ◽  
Active Faults ◽  
Earth’S Crust ◽  
Arabian Plate ◽  
Moho Depth ◽  
Mean Square ◽  
Mean Square Deviation ◽  
Earth's Crust

SUMMARY The study of the discontinuity between crust and mantle beneath Iran is still an open issue in the geophysical community due to its various tectonic features created by the collision between the Iranian and Arabian Plate. For instance in regions such as Zagros, Alborz or Makran, despite the number of studies performed, both by exploiting gravity or seismic data, the depth of the Moho and also interior structure is still highly uncertain. This is due to the complexity of the crust and to the presence of large short wavelength signals in the Moho depth. GOCE observations are capable and useful products to describe the Earth’s crust structure either at the regional or global scale. Furthermore, it is plausible to retrieve important information regarding the structure of the Earth’s crust by combining the GOCE observations with seismic data and considering additional information. In the current study, we used as observation a grid of second radial derivative of the anomalous gravitational potential computed at an altitude of 221 km by means of the space-wise approach, to study the depth of the Moho. The observations have been reduced for the gravitational effects of topography, bathymetry and sediments. The residual gravity has been inverted accordingly to a simple two-layer model. In particular, this guarantees the uniqueness of the solution of the inverse problem which has been regularized by means of a collocation approach in the frequency domain. Although results of this study show a general good agreement with seismically derived depths with a root mean square deviation of 6 km, there are some discrepancies under the Alborz zone and also Oman sea with a root mean square deviation up 10 km for the former and an average difference of 3 km for the latter. Further comparisons with the natural feature of the study area, for instance, active faults, show that the resulting Moho features can be directly associated with geophysical and tectonic blocks.

scholarly journals A Variable Data Fusion Approach for Electromechanical Impedance-Based Damage Detection

Sensors ◽  
2020 ◽  
Vol 20 (15) ◽  
pp. 4204
Author(s):  
Shishir Kumar Singh ◽  
Rohan Soman ◽  
Tomasz Wandowski ◽  
Pawel Malinowski
Keyword(s):  
Data Fusion ◽  
Damage Detection ◽  
Root Mean Square ◽  
Root Mean Square Deviation ◽  
Health Monitoring ◽  
Mean Square ◽  
Mean Square Deviation ◽  
Damage Scenarios ◽  
Electromechanical Impedance ◽  
Fusion Approach

There is continuing research in the area of structural health monitoring (SHM) as it may allow a reduction in maintenance costs as well as lifetime extension. The search for a low-cost health monitoring system that is able to detect small levels of damage is still on-going. The present study is one more step in this direction. This paper describes a data fusion technique by combining the information for robust damage detection using the electromechanical impedance (EMI) method. The EMI method is commonly used for damage detection due to its sensitivity to low levels of damage. In this paper, the information of resistance (R) and conductance (G) is studied in a selected frequency band and a novel data fusion approach is proposed. A novel fused parameter (F) is developed by combining the information from G and R. The difference in the new metric under different damage conditions is then quantified using established indices such as the root mean square deviation (RMSD) index, mean absolute percentage deviation (MAPD), and root mean square deviation using k-th state as the reference (RMSDk). The paper presents an application of the new metric for detection of damage in three structures, namely, a thin aluminum (Al) plate with increasing damage severity (simulated with a drilled hole of increasing size), a glass fiber reinforced polymer (GFRP) composite beam with increasing delamination and another GFRP plate with impact-induced damage scenarios. Based on the experimental results, it is apparent that the variable F increases the robustness of the damage detection as compared to the quantities R and G.

Sign in / Sign up

Export Citation Format

Share Document