Exploring the Impacts of Conformer Selection Methods on Ion Mobility Collision Cross Section Predictions

Abstract

The prediction of structure dependent molecular properties, such as collision cross sections as measured using ion mobility spectrometry, are crucially dependent on the selection of the correct population of molecular conformers. Here, we report an in-depth evaluation of multiple conformation selection techniques, including simple averaging, Boltzmann weighting, lowest energy selection, low energy threshold reductions, and similarity reduction. Generating 50 000 conformers each for 18 molecules, we used the In Silico Chemical Library Engine (ISiCLE) to calculate the collision cross sections for the entire data set. First, we employed Monte Carlo simulations to understand the variability between conformer structures as generated using simulated annealing. Then we employed Monte Carlo simulations to the aforementioned conformer selection techniques applied on the simulated molecular property: the ion mobility collision cross section. Based on our analyses, we found Boltzmann weighting to be a good trade-off between precision and theoretical accuracy. Combining multiple techniques revealed that energy thresholds and root-mean-squared deviation-based similarity reductions can save considerable computational expense while maintaining property prediction accuracy. Molecular dynamic conformer generation tools like AMBER can continue to generate new lowest energy conformers even after tens of thousands of generations, decreasing precision between runs. This reduced precision can be ameliorated and theoretical accuracy increased by running density functional theory geometry optimization on carefully selected conformers.

Similarity Downselection: Finding the n Most Dissimilar Molecular Conformers for Reference-Free Metabolomics

“…Finding the two most dissimilar conformers is trivial, but finding the subsets of three or more is not. As part of a larger analysis to assess the validity of various conformer selection techniques [ 6 ], a method was needed to efficiently find the subset of the n most dissimilar molecular conformers out of a set of 50,000 conformers (where 1 < n < 50,000). These conformers were generated by a modified ISiCLE pipeline on a set of 18 small molecules with mass ranges of ~100–700 Da, as described previously [ 6 ].…”

“…A nontrivial problem is creating sets of conformers and selecting the critical few that should be used for the final property calculations. Thus, several groups have striven to create algorithms and automated pipelines in order to smartly sample conformers sets: e.g., clustering methods [ 1 ], randomly sampling (with or without simulated annealing) [ 2 , 3 , 4 ], or selecting low energy sets from classical or semiempirical molecular modeling [ 5 , 6 ]. Part of the goal of downselecting to a smaller set of conformers is to save on the relatively high computational cost when performing quantum chemical calculations.…”

“…Part of the goal of downselecting to a smaller set of conformers is to save on the relatively high computational cost when performing quantum chemical calculations. In a recent study, we found that by selecting the two most dissimilar and single most similar conformers from simulated annealing runs prior to performing density functional theory calculations, we could save substantial computational time while preserving property calculation accuracy [ 6 ]. However, that method was not scalable, and so we endeavored to find a new subsetting method that could work well in high-throughput pipelines and consider thousands of conformers.…”

Proton Affinity Values of Fentanyl and Fentanyl Analogues Pertinent to Ambient Ionization and Detection

“…One possible explanation for the discrepancy with fentanyl could be that the structure was trapped in a local energy minimum. A way to mitigate this issue is by having more robust conformer sampling, such as generating or testing more structures or by differing the sampling methods . When optimizing the protonated fentanyl conformer structure provided in Lau et al at the same level of theory as the other molecules in Table , the LE BD was 1.630 Å, which is indicative of internal bridging and is more in line with the other bond distances (see Table , fentanyl*).…”

“…Proton affinity and gas-phase basicity values were generated according to eqs 4 and 5 for all amines and amides adducts described in the Validation Library Section. The calculated values for all conformers of a protonated isomer population are Boltzmann weighted 39 according to total energies, yielding one PA and GB for each ion location.…”