Small-Molecule Conformer Generators: Evaluation of Traditional Methods and AI Models on High-Quality Data Sets

Abstract

Small-molecule conformer generation (SMCG) is an extremely important task in both ligand-and structure-based computer-aided drug design, especially during the hit discovery phase. Recently, a multitude of artificial intelligence (AI) models tailored for SMCG have emerged. Despite developers typically furnishing performance evaluation data upon releasing their AI models, a comprehensive and equitable performance comparison between AI models and conventional methods is still lacking. In this study, we curated a new benchmarking data set comprising 3354 high-quality ligand bioactive conformations. Subsequently, we conducted a systematic assessment of the performance of four widely adopted traditional methods (i.e., ConfGenX, Conformator, OMEGA, and RDKit ETKDG) and five AI models (i.e., ConfGF, DMCG, GeoDiff, GeoMol, and torsional diffusion) in the tasks of reproducing bioactive and low-energy conformations of small molecules. In the former task, the AI models have no advantage, particularly with a maximum ensemble size of 1. Even the best-performing AI model GeoMol is still worse than any of the tested traditional methods. Conversely, in the latter task, the torsional diffusion model shows obvious advantages, surpassing the bestperforming traditional method ConfGenX by 26.09 and 12.97% on the COV-R and COV-P metrics, respectively. Furthermore, the influence of force field-based fine-tuning on the quality of the generated conformers was also discussed. Finally, a user-friendly Web server called fastSMCG was developed to enable researchers to rapidly and flexibly generate small-molecule conformers using both traditional and AI methods. We anticipate that our work will offer valuable practical assistance to the scientific community in this field.

Gram matrix: an efficient representation of molecular conformation and learning objective for molecular pretraining

“…To compare the effectiveness of conformation prediction models, we evaluated the performance of our Pre-GTM d model alongside other conformation prediction methods using a test dataset of 3354 high-quality ligand bioactive conformations [ 59 ]. And the results are summarized in Table 6 .…”

“…Motivated by these considerations, we reference the work of Hou et al . [ 59 ] and compare the performance of our model with other conformation generation models based on the platinum diversity benchmark. The methods under comparison include traditional conformation prediction approaches (ConfGenX [ 60 ], Conformator [ 61 ], OMEGA [ 62 ], and RDKit) as well as six AI-based conformation prediction methods (ConfGF [ 63 ], DMCG [ 64 ], GeoDiff [ 65 ], GeoMol [ 58 ], torsional diffusion [ 66 ], and Uni-mol).…”

“…The COV (coverage) and MAT (matching) metrics were utilized to quantify the quality of conformations [ 59 ]. The metrics are defined as:…”

Efficient Generation of Conformer Ensembles Using Internal Coordinates and a Generative Directional Graph Convolution Neural Network

“…Multiple approaches to fast CG have been proposed − (reviewed in ref ), typically based on Monte Carlo (MC) sampling with force-field energy, distance geometry (DG), and rules/knowledge extracted from experimental structures in various combinations. In particular, neural networks (NN), when run on GPUs, may achieve speed advantages over traditional techniques, and several NN-based conformer-generation methods have been reported. − For conformer generation, operating on a roto-translationally invariant description of the molecular 3D structure is desirable. On the other hand, in the standard Cartesian XYZ coordinate description, conformational degrees of freedom are mixed with overall rotations and translations of the molecule.…”