Conformational analysis of fused ring systems using tensor decomposition methods (CROSBI ID 681756)
Prilog sa skupa u zborniku | sažetak izlaganja sa skupa
Podaci o odgovornosti
Sović, Karlo ; Hrenar, Tomica
engleski
Conformational analysis of fused ring systems using tensor decomposition methods
A new procedure for full conformational analysis [1] comprising statistical analysis of molecular dynamics trajectory was applied on fused ring systems like decalin, bicyclo[3.3.0]octane and bicyclo[2.2.0]hexane. This method includes a coordinate space sampling using molecular dynamics simulations with on-the-flycalculation of forces, which serves as a sampling space for locating the initial guess structures for further geometry optimization at a higher level of calculation and conformer clustering. The initial guess structures are extracted from those points in phase space which are strict local maxima in a probability distribution of molecular geometry coordinates. Idea behind this is that during the molecular dynamics run, the investigated molecule would statistically spend more time in and closely around the minima points on the potential energy surface and that, consequently, the probability distribution of the molecular structures in these points of the phase space would have a strict local maxima. Dimensionality of this search is firstly reduced using 2nd-order tensor decomposition tool principal component analysis on the internal coordinate space of investigated molecule. In this reduced space of score values, n-dimensional probability distribution is generated and subjected to a procedure for finding all strict local maxima. Further investigation of this method included multi-way analysis of temperature dependent trajectories for fused ring systems using fixed total number of steps in molecular dynamics runs for comparison of obtained conformational spaces. Molecular dynamics simulations were performed using our own program qcc[2] where the forces were calculated using the PM7method [3] implemented in MOPAC2016 [4] and numerical integration with the Velocity Verlet algorithm. Multi-way analysis of trajectories was performed using our own program moonee[5].
Conformational analysis ; tensor decomposition ; pca ; multi-way analysis
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
2019.
objavljeno
Podaci o matičnoj publikaciji
Computational Chemistry Day 2019 : Book of Abstracts
Zagreb:
978-953-6076-51-2
Podaci o skupu
Computational Chemistry Day 2019
predavanje
11.05.2019-11.05.2019
Zagreb, Hrvatska