Conformer Generation Research Articles

Modelling solution structures of cyclic peptides - How good are we?

Cyclic peptides (CPs) are macrocycles made from amino acids that can bind to classically undruggable drug targets. Controlling pharmacokinetic properties of these macrocyclic peptides is an open question, limited by having precise knowledge of the underlying solution structures. CP solution dynamics are complex: individual structures are often separated by large energy barriers; the presence of conformers is environment dependent and solution conformers can be different from bioactive structures. Specialised simulation methods for CPs have previously been designed to reproduce available x-ray structures, but it is often unknown how well solution structures are reproduced. Here, we report the largest cyclic peptide solution NMR dataset “MacroConf”, which we built from existing literature. We use a range of molecular dynamics methods and specialised cheminformatics methods to model the solution structures of over 30 cyclic peptides from this dataset. We explore (a) how different methods perform relative to one another and to the experimental reference structures and (b) whether computationally cheaper methods can reach the accuracy of gold-standard accelerated molecular dynamics simulations to allow for fast prediction of solution ensembles. By providing automated computational workflows, this work is easily extendable, fully reproducible, and is built as a reference point to guide development of future conformer generators for cyclic peptides.

Advances in deep learning-based 3D molecular generative models

<p indent="0mm">Molecular generation, aiming at designing novel molecules with desired properties such as biological activity and drug metabolism and pharmacokinetic (DMPK) properties in a low-cost and high-efficient manner, is a fundamental problem in drug discovery. Recently, deep generative models have been widely used in drug discovery, along with numerous model architectures and optimization strategies being explored, most of which are developed to generate one-dimensional or two-dimensional molecular structures. Inspired by the rapid development of deep learning in processing geometric graph data, 3D generative models for molecular generation have also been proposed, gaining attention for their advantages and potential on direct 3D molecular conformation generation and structure-based drug discovery. This survey has offered a systematic summarization of existing research achievements of the domestic and foreign researchers in recent years in the aspects of 3D molecular generation. According to the input of 3D molecular generation algorithms, it is divided into latent variable-based generation, 2D graph-based generation and 3D geometry-based generation. Then, according to the output of 3D molecular generation algorithms, it is divided into goal-directed generation and non-goal-directed generation. Furthermore, the performance of different generative models on main public datasets is summarized, which proves the advantages and disadvantages of the various models. Finally, some promising research directions are proposed.

SDEGen: learning to evolve molecular conformations from thermodynamic noise for conformation generation.

Generation of representative conformations for small molecules is a fundamental task in cheminformatics and computer-aided drug discovery, but capturing the complex distribution of conformations that contains multiple low energy minima is still a great challenge. Deep generative modeling, aiming to learn complex data distributions, is a promising approach to tackle the conformation generation problem. Here, inspired by stochastic dynamics and recent advances in generative modeling, we developed SDEGen, a novel conformation generation model based on stochastic differential equations. Compared with existing conformation generation methods, it enjoys the following advantages: (1) high model capacity to capture multimodal conformation distribution, thereby searching for multiple low-energy conformations of a molecule quickly, (2) higher conformation generation efficiency, almost ten times faster than the state-of-the-art score-based model, ConfGF, and (3) a clear physical interpretation to learn how a molecule evolves in a stochastic dynamics system starting from noise and eventually relaxing to the conformation that falls in low energy minima. Extensive experiments demonstrate that SDEGen has surpassed existing methods in different tasks for conformation generation, interatomic distance distribution prediction, and thermodynamic property estimation, showing great potential for real-world applications.

A flexible data-free framework for structure-based de novo drug design with reinforcement learning.

Contemporary structure-based molecular generative methods have demonstrated their potential to model the geometric and energetic complementarity between ligands and receptors, thereby facilitating the design of molecules with favorable binding affinity and target specificity. Despite the introduction of deep generative models for molecular generation, the atom-wise generation paradigm that partially contradicts chemical intuition limits the validity and synthetic accessibility of the generated molecules. Additionally, the dependence of deep learning models on large-scale structural data has hindered their adaptability across different targets. To overcome these challenges, we present a novel search-based framework, 3D-MCTS, for structure-based de novo drug design. Distinct from prevailing atom-centric methods, 3D-MCTS employs a fragment-based molecular editing strategy. The fragments decomposed from small-molecule drugs are recombined under predefined retrosynthetic rules, offering improved drug-likeness and synthesizability, overcoming the inherent limitations of atom-based approaches. Leveraging multi-threaded parallel simulations combined with a real-time energy constraint-based pruning strategy, 3D-MCTS achieves remarkable efficiency. At a fixed computational cost, it outperforms other state-of-the-art (SOTA) methods by producing molecules with enhanced binding affinity. Furthermore, its fragment-based approach ensures the generation of more dependable binding conformations, exhibiting a success rate 43.6% higher than that of other SOTAs. This advantage becomes even more pronounced when handling targets that significantly deviate from the training dataset. 3D-MCTS is capable of achieving thirty times more hits with high binding affinity than traditional virtual screening methods, which demonstrates the superior ability of 3D-MCTS to explore chemical space. Moreover, the flexibility of our framework makes it easy to incorporate domain knowledge during the process, thereby enabling the generation of molecules with desirable pharmacophores and enhanced binding affinity. The adaptability of 3D-MCTS is further showcased in metalloprotein applications, highlighting its potential across various drug design scenarios.

Relaxation dynamics of high-energy excited states of carotenoids studied by UV excitation and pump-repump-probe transient absorption spectroscopy.

The excited states of carotenoids have been a subject of numerous studies. While a majority of these reports target the excited state dynamics initiated by the excitation of the S2 state, the upper excited state(s) absorbing in the UV spectral region (denoted as SUV) has been only scarcely studied. Moreover, the relation between the SUV and Sn, the final state of the well-known S1-Sn transition of carotenoids, remains unknown. To address this yet-unresolved issue, we compared the excited state dynamics of two carotenoids, namely, β-carotene and astaxanthin, after excitation of either the SUV or Sn state. The SUV state was excited directly by UV light, and the excitation of the Sn state was achieved via re-pumping the S1-Sn transition. The results indicated that direct SUV excitation produces an S1-Sn band that is significantly broader than that obtained after S2 excitation, most probably due to the generation of multiple S1 conformations produced by excess energy. No such broadening is observed if the Sn state is excited by the re-pump pulse. This shows that the Sn and SUV states are different, each initializing a specific relaxation pathway. We propose that the Sn state retains the coupled triplet pair character of the S1 state, while the SUV state is the higher state of Bu+ symmetry accessible by one-photon transition.

Mapping the energy landscape of PROTAC-mediated protein-protein interactions

A principal challenge in computational modeling of macromolecules is the vast conformational space that arises out of large numbers of atomic degrees of freedom. Recently, growing interest in building predictive models of complexes mediated by Proteolysis Targeting Chimeras (PROTACs) has led to the application of state-of-the-art computational techniques to tackle this problem. However, repurposing existing tools to carry out protein-protein docking and linker conformer generation independently results in extensive sampling of structures incompatible with PROTAC-mediated complex formation. Here we show that it is possible to restrict the search to the space of protein-protein conformations that can be bridged by a PROTAC molecule with a given linker composition by using a cyclic coordinate descent algorithm to position PROTACs into complex-bound configurations. We use this methodology to construct potential energy and solvation energy landscapes of PROTAC-mediated interactions. Our results suggest that desolvation of amino acids at interfaces could play a dominant role in PROTAC-mediated complex formation.

3D Conformational Generative Models for Biological Structures Using Graph Information-Embedded Relative Coordinates.

Developing molecular generative models for directly generating 3D conformation has recently become a hot research area. Here, an autoencoder based generative model was proposed for molecular conformation generation. A unique feature of our method is that the graph information embedded relative coordinate (GIE-RC), satisfying translation and rotation invariance, was proposed as a novel way for encoding molecular three-dimensional structure. Compared with commonly used Cartesian coordinate and internal coordinate, GIE-RC is less sensitive on errors when decoding latent variables to 3D coordinates. By using this method, a complex 3D generation task can be turned into a graph node feature generation problem. Examples were shown that the GIE-RC based autoencoder model can be used for both ligand and peptide conformation generation. Additionally, this model was used as an efficient conformation sampling method to augment conformation data needed in the construction of neural network-based force field.

Enhanced soft limits in de Sitter space

In flat space, the scattering amplitudes of certain scalar effective field theories exhibit enhanced soft limits due to the presence of hidden symmetries. In this paper, we show that this phenomenon extends to wavefunction coefficients in de Sitter space. Using a representation in terms of boundary conformal generators acting on contact diagrams, we find that imposing enhanced soft limits fixes the masses and four-point couplings (including curvature corrections) in agreement with Lagrangians recently derived from hidden symmetries. Higher-point couplings can then be fixed using a bootstrap procedure which we illustrate at six points. We also discuss implications for the double copy in de Sitter space.

Graph-Convolutional Neural Net Model of the Statistical Torsion Profiles for Small Organic Molecules.

We present a graph-convolutional neural network (GCNN)-based method for learning and prediction of statistical torsional profiles (STP) in small organic molecules based on the experimental X-ray structure data. A specialized GCNN torsion profile model is trained using the structures in the Crystallography Open Database (COD). The GCNN-STP model captures torsional preferences over a wide range of torsion rotor chemotypes and correctly predicts a variety of effects from the vicinal atoms and moieties. GCNN-STP statistical profiles also show good agreement with quantum chemically (DFT) calculated torsion energy profiles. Furthermore, we demonstrate the application of the GCNN-STP statistical profiles for conformer generation. A web server that allows interactive profile prediction and viewing is made freely available at https://www.molsoft.com/tortool.html.

Holo Protein Conformation Generation from Apo Structures by Ligand Binding Site Refinement.

An important part in structure-based drug design is the selection of an appropriate protein structure. It has been revealed that a holo protein structure that contains a well-defined binding site is a much better choice than an apo structure in structure-based drug discovery. Therefore, it is valuable to obtain a holo-like protein conformation from apo structures in the case where no holo structure is available. Meeting the need, we present a robust approach to generate reliable holo-like structures from apo structures by ligand binding site refinement with restraints derived from holo templates with low homology. Our method was tested on a test set of 32 proteins from the DUD-E data set and compared with other approaches. It was shown that our method successfully refined the apo structures toward the corresponding holo conformations for 23 of 32 proteins, reducing the average all-heavy-atom RMSD of binding site residues by 0.48 Å. In addition, when evaluated against all the holo structures in the protein data bank, our method can improve the binding site RMSD for 14 of 19 cases that experience significant conformational changes. Furthermore, our refined structures also demonstrate their advantages over the apo structures in ligand binding mode predictions by both rigid docking and flexible docking and in virtual screening on the database of active and decoy ligands from the DUD-E. These results indicate that our method is effective in recovering holo-like conformations and will be valuable in structure-based drug discovery.

Stress-inducible phosphoprotein 1 (HOP/STI1/STIP1) regulates the accumulation and toxicity of α-synuclein in vivo

The predominantly pre-synaptic intrinsically disordered protein α-synuclein is prone to misfolding and aggregation in synucleinopathies, such as Parkinson’s disease (PD) and Dementia with Lewy bodies (DLB). Molecular chaperones play important roles in protein misfolding diseases and members of the chaperone machinery are often deposited in Lewy bodies. Here, we show that the Hsp90 co-chaperone STI1 co-immunoprecipitated α-synuclein, and co-deposited with Hsp90 and Hsp70 in insoluble protein fractions in two mouse models of α-synuclein misfolding. STI1 and Hsp90 also co-localized extensively with filamentous S129 phosphorylated α-synuclein in ubiquitin-positive inclusions. In PD human brains, STI1 transcripts were increased, and in neurologically healthy brains, STI1 and α-synuclein transcripts correlated. Nuclear Magnetic Resonance (NMR) analyses revealed direct interaction of α-synuclein with STI1 and indicated that the STI1 TPR2A, but not TPR1 or TPR2B domains, interacted with the C-terminal domain of α-synuclein. In vitro, the STI1 TPR2A domain facilitated S129 phosphorylation by Polo-like kinase 3. Moreover, mice over-expressing STI1 and Hsp90ß presented elevated α-synuclein S129 phosphorylation accompanied by inclusions when injected with α-synuclein pre-formed fibrils. In contrast, reduced STI1 function decreased protein inclusion formation, S129 α-synuclein phosphorylation, while mitigating motor and cognitive deficits as well as mesoscopic brain atrophy in α-synuclein-over-expressing mice. Our findings reveal a vicious cycle in which STI1 facilitates the generation and accumulation of toxic α-synuclein conformers, while α-synuclein-induced proteostatic stress increased insoluble STI1 and Hsp90.

Benchmark of 3D conformer generation and molecular property calculation for medium-sized molecules

Benchmark of 3D conformer generation and molecular property calculation for medium-sized molecules

Conformer-RL: A deep reinforcement learning library for conformer generation.

Conformer‐RL is an open‐source Python package for applying deep reinforcement learning (RL) to the task of generating a diverse set of low‐energy conformations for a single molecule. The library features a simple interface to train a deep RL conformer generation model on any covalently bonded molecule or polymer, including most drug‐like molecules. Under the hood, it implements state‐of‐the‐art RL algorithms and graph neural network architectures tuned specifically for molecular structures. Conformer‐RL is also a platform for researching new algorithms and neural network architectures for conformer generation, as the library contains modular class interfaces for RL environments and agents, allowing users to easily swap components with their own implementations. Additionally, it comes with tools to visualize and save generated conformers for further analysis. Conformer‐RL is well‐tested and thoroughly documented with tutorials for each of the functionalities mentioned above, and is available on PyPi and Github: https://github.com/ZimmermanGroup/conformer-rl.

Conformal bridge transformation, mathcal{PT} - and supersymmetry

Supersymmetric extensions of the 1D and 2D Swanson models are investigated by applying the conformal bridge transformation (CBT) to the first order Berry-Keating Hamiltonian multiplied by i and its conformally neutral enlargements. The CBT plays the role of the Dyson map that transforms the models into supersymmetric generalizations of the 1D and 2D harmonic oscillator systems, allowing us to define pseudo-Hermitian conjugation and a suitable inner product. In the 1D case, we construct a mathcal{PT} -invariant supersymmetric model with N subsystems by using the conformal generators of supersymmetric free particle, and identify its complete set of the true bosonic and fermionic integrals of motion. We also investigate an exotic N = 2 supersymmetric generalization, in which the higher order supercharges generate nonlinear superalgebras. We generalize the construction for the 2D case to obtain the mathcal{PT} -invariant supersymmetric systems that transform into the spin-1/2 Landau problem with and without an additional Aharonov-Bohm flux, where in the latter case, the well-defined integrals of motion appear only when the flux is quantized. We also build a 2D supersymmetric Hamiltonian related to the “exotic rotational invariant harmonic oscillator” system governed by a dynamical parameter γ. The bosonic and fermionic hidden symmetries for this model are shown to exist for rational values of γ.

An optimization algorithm for conformer generation based on the bond contribution ranking

Many works in computational drug discovery require the conformer generation of small molecules. Most existing tools aim to generate diverse conformers and deal with all of the rotatable bonds without distinction. There are some problems in existing approaches, such as the combinatorial explosion of conformers along with the number of rotatable bonds increasing and the incomplete sampling of the conformational space. Here, we present an optimized conformer generation algorithm based on the bond contribution ranking (ABCR) to find the optimal conformer under any specified scoring function.Compared with existing methods, our method can improve molecular conformational searching and protein-ligand docking performance. Meanwhile, our method has the same or broader coverage of conformational space in the global conformer sampling. Our research shows it can achieve the optima with small numbers of generated conformers and small numbers of iterations.

Effective field theories and cosmological scattering equations

We propose worldsheet formulae for wavefunction coefficients of the massive non-linear sigma model (NLSM), scalar Dirac-Born-Infeld (DBI), and special Galileon (sGal) theories in de Sitter momentum space in terms of the recently proposed cosmological scattering equations constructed from conformal generators in the future boundary. The four-point integrands are assembled from simple building blocks and we identify a double copy prescription mapping the NLSM wavefunction coefficient to the DBI and sGal wavefunction coefficients, including mass deformations and curvature corrections. Finally, we compute the soft limits of these wavefunction coefficients and find that they can be written in terms of boundary conformal generators acting on contact diagrams.

Indexing Ultrafast Shape-Based Descriptors in MongoDB to Identify TLR4 Pathway Agonists.

A method is presented for an ultrafast shape-based search workflow for the screening of large compound collections, i.e., those of vendors. The three-dimensional shape of a molecule dictates its biological activity by enabling the molecule to fit into binding pockets of proteins. Quite often, distinctly different chemical compounds that have similar shapes can bind in a similar way. OpenEye pioneered an algorithm for comparing shapes of molecules by overlaying them in a computer and measuring differences between a query molecule and a target molecule. Overlaying shapes is a computationally intensive process and represents a bottleneck in searching for similar molecules. More recent publications describe alternative methods of overlaying molecules, which are accomplished by comparing shape-based descriptors. These methods were implemented in the Open Drug Discovery Toolkit (ODDT) package. We utilized a combination of open-source software packages like ODDT and RDkit to implement a workflow for ultrafast conformer generation and matching that does not require storing precomputed conformers on the file system or in memory. Moreover, the generated descriptors could be optionally stored in MongoDB for performing searches in the future. To speed up the search, we created a set of indexes from the transformed shape-based descriptors. We are in the process of calculating descriptors for multiple vendors, including Enamine's "REAL" collection of 1.2 billion compounds. Currently, the shape similarity search on more than 70 million compounds takes less than 8 s! We exemplified our methodology with the screen of compounds that can act as putative TLR4 agonists. The search was based on a literature-known small-molecule TLR4 agonist series. In due course, we identified compounds with novel structural motifs that were active in mouse and human TLR4 reporter cell lines.

On the differential representation and color-kinematics duality of AdS boundary correlators

The AdS boundary correlators and their dual correlation functions of boundary operators have been the main dynamic observables of the holographic duality relating a bulk AdS theory and a boundary conformal field theory. We show that tree-level AdS boundary correlators for generic states can be expressed as nonlocal differential operators of a certain structure acting on contact Witten diagrams. We further write the boundary correlators in a form that is very similar to flat space amplitudes, with Mandelstam variables replaced by certain combinations of single-state conformal generators, prove that all tree-level AdS boundary correlators have a differential representation, and detail the conversion of such differential expressions to position space. We illustrate the construction through the computation of the boundary correlators of scalars coupled to gluons and gravitons; when converted to position space, they reproduce known results. Color-kinematics duality and BCJ relations can be defined in analogy with their flat space counterparts, and are respected by the scalar correlators with a gluon exchange. We also discuss potential approaches to the double copy and find that its direct generalization may require nontrivial extensions.

Efficient 3D conformer generation of cyclic peptides formed by a disulfide bond

Cyclic peptides formed by disulfide bonds have been one large group of common drug candidates in drug development. Structural information of a peptide is essential to understand its interaction with its target. However, due to the high flexibility of peptides, it is difficult to sample the near-native conformations of a peptide. Here, we have developed an extended version of our MODPEP approach, named MODPEP2.0, to fast generate the conformations of cyclic peptides formed by a disulfide bond. MODPEP2.0 builds the three-dimensional (3D) structures of a cyclic peptide from scratch by assembling amino acids one by one onto the cyclic fragment based on the constructed rotamer and cyclic backbone libraries. Being tested on a data set of 193 diverse cyclic peptides, MODPEP2.0 obtained a considerable advantage in both accuracy and computational efficiency, compared with other sampling algorithms including PEP-FOLD, ETKDG, and modified ETKDG (mETKDG). MODPEP2.0 achieved a high sampling accuracy with an average Calpha RMSD of 2.20 Å and 1.66 Å when 10 and 100 conformations were considered, respectively, compared with 3.41 Å and 2.62 Å for PEP-FOLD, 3.44 Å and 3.16 Å for ETKDG, 3.09 Å and 2.72 Å for mETKDG. MODPEP2.0 also reproduced experimental peptide structures for 81.35% of the test cases when an ensemble of 100 conformations were considered, compared with 54.95%, 37.50% and 50.00% for PEP-FOLD, ETKDG, and mETKDG. MODPEP2.0 is computationally efficient and can generate 100 peptide conformations in one second. MODPEP2.0 will be useful in sampling cyclic peptide structures and modeling related protein-peptide interactions, facilitating the development of cyclic peptide drugs.

Marine macrolides as an efficient source of FMS-like tyrosine kinase 3 inhibitors: A comprehensive approach of in silico virtual screening

Marine organisms are a definitive source of antibiotics and kinase inhibitors which provide cues for discovering novel drug leads. Marine macrolides are getting much attraction due to their enzyme inhibitory potential. The present study comprehensively dealt with the virtual screening and structure-based prediction of macrolide compounds against FMS-like tyrosine kinase 3 receptors (FLT3). The FLT3 was chosen as a biological target against the 990 marine macrolides. Before the virtual screening of macrolide compounds, validation of molecular docking was carried out by re-docking of co-crystallized Gilteritinib within the FLT3. Among the selected 990 candidates of marine macrolides, 311 were failed due to the generation of insufficient conformers. Amongst the successful compounds, 22 compounds were also failed to dock within the receptor, while the remaining 657 marine macrolide entities elicited successful docking. The HYBRID Chemguass4 Score ranged from -10.17 to -0.02. This vast difference in the HYBRID ChemGuass4 score is attributed to the difference in binding potential with the receptor's binding pocket. The top ten compounds were selected based on the HYBRID ChemGuass4 Score lower than -8.0 against FLT3. The pharmacokinetics and ADME properties revealed the drug likeliness of the macrolides.