ChemRxiv
Not a member yet
27047 research outputs found
Sort by
Guided docking as a data generation approach facilitates structure-based machine learning on kinases
No full textDrug discovery pipelines nowadays rely on machine learning models to explore and evaluate large chemical spaces. While including 3D structural information is considered beneficial, structural models are hindered by the availability of protein-ligand complex structures. Exemplified for kinase drug discovery, we address this issue by generating kinase-ligand complex data using template docking for the kinase compound subset of available ChEMBL assay data. To evaluate the benefit of the created complex data, we use it to train a structure-based E(3)-invariant graph neural network (GNN). Our evaluation shows that binding affinities can be predicted with significantly higher precision by models that take synthetic binding poses into account compared to ligand or DTI models only
Modular soft stretchable low-cost elastomers for stereolithography printing structures with extreme dissipative properties
No full textAdditive manufacturing of elastomers enables the fabrication of many technologically important structures and devices. However, it remains a challenge to develop soft and stretchable elastomers for stereolithography (SLA) printing, one of the most used additive manufacturing techniques for producing objects with relatively high-resolution and smooth finishes. Here, we report a modular, soft, stretchable, and low-cost elastomer resin for SLA printing. The resin consists of mainly commodity acrylates and can be photocured to form a dual-crosslinked network containing covalent and reversible crosslinks. Controlling the ratio of covalent and reversible crosslinks, we create elastomers with an exceptional combination of softness and stretchability (Young’s modulus of 20-150 kPa and tensile breaking strain of 510-1350%) that cannot be achieved by existing SLA resins. We demonstrate printing this resin to produce high-resolution three-dimensional (3D) structures with extreme dissipative properties. Further, we develop a setup to show that the 3D structures can protect brain-like soft gels from impact damage in reducing the severity of impact by 75%. Together with the low-cost of raw chemicals and modular nature of the design, our soft and stretchable elastomer resins provide a new class of soft materials for high-fidelity additive manufacturing of functional architectures
Tuning and Matching Error Compensated, Quantitative Solid-state NMR
No full textAbstract NMR spectroscopy has long been recognized as a powerful quantitative analytical tool. Quantification is commonly done against internal and external standards. A third approach is to quantify against an electronic reference which combines the advantages of the two methods. The implementation of this approach in solid-state NMR is more challenging due to the single-coil design of double resonance probes. In this study, a novel approach for implementing the electronic referencing method in solid-state NMR by injecting the reference signal using a broadband antenna installed near the NMR receiver coil is presented. This method demonstrates excellent accuracy and precision, as it remains robust to changes in the electronic conditions of the probe, including tuning and matching errors
Exceptions, Paradoxes, and Their Resolutions in Chemical Reactivities
No full textChemists have long utilized theories and models to rationalize empirical data and predict outcomes in new chemical systems. Unexpected reactivities often present themselves as exceptions or paradoxes, highlighting gaps in our current understanding and the limitations of (then) existing models. Rather than something to be disregarded, these exceptions offer valuable insights that can lead to the refinement of theories and the resolution of paradoxes, ultimately fostering scientific discovery. While the concept of scientific discovery described is not new, practically, it remains significantly underappreciated in the field of organic chemistry. This perspective article postulates possible epistemic origins of such unexpected chemical reactivities across various contexts, proposing a systematic approach to addressing these anomalies. Through a compilation of origins, supported by specific modern examples, this work aims to lay the groundwork for more comprehensive thinking that will help the field navigate future exploration. While the illustrative examples presented herein focus on physical organic chemistry, the concepts discussed are universally applicable across different fields in chemical science
Digitally Enabled Generic Analytical Framework Accelerating the Pace of Liquid Chromatography Method Development for Vaccine Adjuvant Formulations
No full textThe growing use of adjuvants in the fast-paced formulation of new vaccines has created an unprecedented need for meaningful analytical assays that deliver reliable quantitative data from complex adjuvant and adjuvant-antigen mixtures. Due to their complex chemical and physical properties, method development for the separation of vaccine adjuvants is considered a highly challenging and laborious task. Reversed-phase liquid chromatography (RPLC) is among the most important tests in the (bio)pharmaceutical industry for release and stability indicating measurements including adjuvant content, identity, and purity profile. However, the time constraints of developing “on-demand” robust quantitative methods prior to each change in formulation can easily lead to sample analysis becoming a bottleneck in vaccine development. Herein a simple and efficient generic analytical framework capable of chromatographically resolving the most commonly used non-aluminum based adjuvants across academic and industrial sectors is introduced. This was designed to seek a more proactive approach for fast-paced assay development endeavors that evolved from extensive stationary phase screening in conjunction with multifactorial in silico simulations of adjuvant retention time (RT) as a function of gradient time, temperature, organic modifier blending, and buffer concentration. The multifactorial retention models yield 3D resolution maps with excellent baseline separation of all adjuvants in a single run, which was found to be very accurate, with differences between experimental and simulated retention times of less than 1%. The analytical framework described here also includes the introduction of a more versatile approach to method development by introducing a dynamic RT database for adjuvants covering the entire library of adjuvants with broad mechanisms of action across numerous vaccine formulations with excellent linearity, accuracy, precision, and specificity. The power of this framework was also demonstrated with numerous analytical assays that can be generated rapidly from simulations guiding vaccine processes in the development of new adjuvant formulations. Analytical assay in this work covers content, purity profile by RPLC-UV-CAD, and component identification (RPLC-MS) across complex vaccine formulations, including the use of surfactants (e.g., polysorbates), as well as their separation from adjuvant targets
Graph-Text Contrastive Learning of Inorganic Crystal Structure toward a Foundation Model of Inorganic Materials
No full textDeveloping foundation models for materials science has attracted attention. However, there is a lack of studies on inorganic materials due to the difficulty in the comprehensive representation of geometric concepts composing crystals: local atomic environments, their connections, and the global symmetries. We present a contrastive learning of inorganic crystal structure (CLICS) for embedding the geometric concepts, which contrasts texts representing the contextual patterns of geometries with the crystal graphs. We demonstrate that the geometric concepts are integrally embedded on the CLICS feature space, through experiments of concept retrieval from crystal graphs, similar structure search, and few-shot/imbalanced crystal structure classification
The effects of solid acids as cocatalysts on the chelation-assisted hydroacylation of alkenes and alkynes
No full textThe use of homogeneous Bronsted acid cocatalysts (such as benzoic acid) in hydroacylation reactions via imine intermediates has been extensively studied, but the use of heterogeneous cocatalysts has been limited to montmorillonite K10. Thus, we can use other solid acids to increase the efficiency of the reaction. Here, we described the effects of sulfated zirconia, Al-MCM-41 or superacid modified montmorillonite on the hydroacylation of alkenes and alkynes with aldehydes via imine intermediates and in the presence of the Wilkinson complex. Furthermore, we addressed the dual role of montmorillonite, a redox reagent in the presence of TEMPO and an acid solid, allowing the direct use of benzyl alcohols as substrates to generate saturated or a-b unsaturated ketones
Employing Metadynamics to Predict the Membrane Partitioning of Carboxy-2H-Azirine Natural Products
No full textNatural products have diverse chemical structures and biological activities which often serve as sources of new therapeutic agents. Those containing a carboxy-2H-azirine moiety are an exciting target for investigation due in part to the broad-spectrum antimicrobial activity these compounds have and the significant chemical space for novel therapeutic development offered by this unique scaffold. The carboxy-2H-azirine moiety, including those appended to well-characterized chemical scaffolds, is understudied, which creates a challenge for understanding potential modes of inhibition. In particular, some known natural product carboxy-2H-azirines have long hydrophobic tails, which might lead to amphipathicity and implicate them in membrane associated processes. Metadynamics is an effective method for calculating the free energy changes associated with membrane embedding processes. In this study, we examined a small set of carboxy-2H-azirine natural products, including analogs with long alkyl chains, geometric isomers, and one comprising the simple carboxy-2H-azirine core. We compared the physiochemical properties of these compounds to those of established membrane embedders with similar chemical scaffolds. This was intended to isolate the physiochemical properties of the carboxy-2H-azirine group and understand molecular influences of this moiety on membrane partitioning. To accomplish this, we developed a force field for the 2H-azirine functional group and performed metadynamics simulations of the partitioning into a model membrane (75 % POPE, 25 % POPG) from aqueous solution. We determine that the carboxy-2H-azirine functional group is likely hydrophilic, imbuing the long chain analogs with amphipathicity similar to the known membrane binding molecules to which they were compared. For the long chain analogs, the carboxy-2H-azirine headgroup stays within 1 nm of the phosphate layer, while the carboxy-2H-azirines lacking the long alkyl chain partitions completely into aqueous solution
Molecular analysis and design using multimodal generative artificial intelligence via multi-agent modeling
No full textWe report the use of a multimodal generative artificial intelligence framework, the X-LoRA-Gemma large language model (LLM), to analyze, design and test molecular design. The X-LoRA-Gemma model, inspired by biological principles and featuring ~7 billion parameters, dynamically reconfigures its structure through a dual-pass inference strategy to enhance its problem-solving abilities across diverse scientific domains. The model is used to first identify molecular engineering targets through a systematic human-AI and AI-AI self-driving multi-agent approach to elucidate key targets for molecular optimization to improve interactions between molecules. Next, a multi-agent generative design process is used that includes rational steps, reasoning and autonomous knowledge extraction. Target properties of the molecule are identified either using a Principal Component Analysis (PCA) of key molecular properties or sampling from the distribution of known molecular properties. The model is then used to generate a large set of candidate molecules, which are analyzed via their molecular structure, charge distribution, and other features. We validate that as predicted, increased dipole moment and polarizability is indeed achieved in the designed molecules. We anticipate an increasing integration of these techniques into the molecular engineering workflow, ultimately enabling the development of innovative solutions to address a wide range of societal challenges. We conclude with a critical discussion of challenges and opportunities of the use of multimodal generative AI for molecular engineering, analysis and design
Can Deep Learning Search for Exceptional Chiroptical Properties? The Halogenated [6]Helicene Case
No full textIn this work we predict, among more than a billion possibilities, the best candidates of halogenated [6]helicenes in order to obtain excellent chiroptical properties in terms of the rotatory strength (R). We have used DFT calculations to randomly create derivatives from 1 to 16 halogen atoms, that were then used as a data set to train different deep neural network models. It is worth noting that the simplest model affords a parametrization that allows to easily predict the value of R for any hexahalogenated [6]helicene. The correlation between calculated and predicted data increases together with the complexity of the model. The results show that some positions and halogens are preferred to increase the R value. In this sense, we have also synthesized the derivatives with the higher predicted R, obtaining excellent correlation among the values obtained experimentally, by DFT-calculations and machine learning predictions