Small molecules deliver big
Related Articles
A data-driven generative strategy to avoid reward hacking in multi-objective molecular design
Molecular design using data-driven generative models has emerged as a promising technology, impacting various fields such as drug discovery and the development of functional materials. However, this approach is often susceptible to optimization failure due to reward hacking, where prediction models fail to extrapolate, i.e., fail to accurately predict properties for designed molecules that considerably deviate from the training data. While methods for estimating prediction reliability, such as the applicability domain (AD), have been used for mitigating reward hacking, multi-objective optimization makes it challenging. The difficulty arises from the need to determine in advance whether the multiple ADs with some reliability levels overlap in chemical space, and to appropriately adjust the reliability levels for each property prediction. Herein, we propose a reliable design framework to perform multi-objective optimization using generative models while preventing reward hacking. To demonstrate the effectiveness of the proposed framework, we designed candidates for anticancer drugs as a typical example of multi-objective optimization. We successfully designed molecules with high predicted values and reliabilities, including an approved drug. In addition, the reliability levels can be automatically adjusted according to the property prioritization specified by the user without any detailed settings.
Imaging molecular structures and interactions by enhanced confinement effect in electron microscopy
Atomic imaging of molecules and intermolecular interactions are of great significance for a deeper understanding of the basic physics and chemistry in various applications, but it is still challenging in electron microscopy due to their thermal mobility and beam sensitivity. Confinement effect and low-dose imaging method may efficiently help us achieve stable high-resolution resolving of molecules and their interactions. Here, we propose a general strategy to image the confined molecules and evaluate the strengths of host-guest interactions in three material systems by low-dose electron microscopy. Then, we change the guest molecules to analyze how each kind of interaction strength influences the imaging quality of these molecules by using a same parameter, the aspect ratios of imaged molecular projections. In the material systems of perovskites (ionic) and zeolites with adsorbed molecules (van der Waals), we can obtain a clear image of molecular configurations by enhancing host-guest interactions. Even in metal organic framework (coordination) system, the atomic structures and bonds of aromatics can be achieved. These results provide a general description on the relation between molecular images and interactions, making it possible to study more molecular behaviors in wide applications by real-space imaging.
The risk effects of corporate digitalization: exacerbate or mitigate?
This study elaborates on the risk effects of corporate digital transformation (CDT). Using the ratio of added value of digital assets to total intangible assets as a measure of CDT, this study overall reveals an inverse relationship between CDT and revenue volatility, even after employing a range of technical techniques to address potential endogeneity. Heterogeneity analysis highlights that the firms with small size, high capital intensity, and high agency costs benefit more from CDT. It also reveals that advancing information infrastructure, intellectual property protection, and digital taxation enhances the effectiveness of CDT. Mechanism analysis uncovers that CDT not only enhances financial advantages such as bolstering core business and mitigating non-business risks but also fosters non-financial advantages like improving corporate governance and ESG performance. Further inquiries into the side effects of CDT and the dynamics of revenue volatility indicate that CDT might compromise cash flow availability. Excessive digital investments exacerbate operating risks. Importantly, the reduction in operating risk associated with CDT does not sacrifice the potential for enhanced company performance; rather, it appears to augment the value of real options.
Molecular optimization using a conditional transformer for reaction-aware compound exploration with reinforcement learning
Designing molecules with desirable properties is a critical endeavor in drug discovery. Because of recent advances in deep learning, molecular generative models have been developed. However, the existing compound exploration models often disregard the important issue of ensuring the feasibility of organic synthesis. To address this issue, we propose TRACER, which is a framework that integrates the optimization of molecular property optimization with synthetic pathway generation. The model can predict the product derived from a given reactant via a conditional transformer under the constraints of a reaction type. The molecular optimization results of an activity prediction model targeting DRD2, AKT1, and CXCR4 revealed that TRACER effectively generated compounds with high scores. The transformer model, which recognizes the entire structures, captures the complexity of the organic synthesis and enables its navigation in a vast chemical space while considering real-world reactivity constraints.
Foundation dentists’ attitudes and experiences in providing dental care for dependant older adults resident in care home settings
There is a continued increase of older dependant adults in England. Foundation Dentists (FDs) are often the dental workforce being tasked with providing dental care to dependant older adults resident in care home settings. This study explores whether FDs have the experience and confidence to deliver this.
Responses