Synthesize heterogeneous biological knowledge via representation learning for Alzheimer's disease drug repurposing

Developing drugs for treating Alzheimer's disease has been extremely challenging and costly due to limited knowledge of underlying mechanisms and therapeutic targets. To address the challenge in AD drug development, we developed a multi-task deep learning pipeline that learns biological interactions and AD risk genes, then utilizes multi-level evidence on drug efficacy to identify repurposable drug candidates. Using the embedding derived from the model, we ranked drug candidates based on evidence from post-treatment transcriptomic patterns, efficacy in preclinical models, population-based treatment effects, and clinical trials. We mechanistically validated the top-ranked candidates in neuronal cells, identifying drug combinations with efficacy in reducing oxidative stress and safety in maintaining neuronal viability and morphology. Our neuronal response experiments confirmed several biologically efficacious drug combinations. This pipeline showed that harmonizing heterogeneous and complementary data/knowledge, including human interactome, transcriptome patterns, experimental efficacy, and real-world patient data shed light on the drug development of complex diseases.

Is the largest aqueous gold cluster a superatom complex? Electronic structure & optical response of the structurally determined Au146(p-MBA)57

The new water-soluble gold cluster Au₁₄₆(p-MBA)₅₇, the structure of which has been recently determined at sub-atomic resolution by Vergara et al., is the largest aqueous gold cluster ever structurally determined and likewise the smallest cluster with a stacking fault. The core presents a twinned truncated octahedron, while additional peripheral gold atoms follow a C₂ rotational symmetry. According to the usual counting rules of the superatom complex (SAC) model, the compound attains a number of 92 SAC electrons if the overall net charge is 3- (three additional electrons). As this is the number of electrons required for a major shell closing, the question arises of whether Au₁₄₆(p-MBA)₅₇ should be regarded as a superatom complex. Starting from the experimental coordinates we have analyzed the structure using density-functional theory. The optimized (relaxed) structure retains all the connectivity of the experimental coordinates, while removing much of its irregularities in interatomic distances, thereby enhancing the C₂-symmetry feature. On analyzing the angular-momentum-projected states, we show that, despite a small gap, the electronic structure does not exhibit SAC model character. In addition, optical absorption spectra are found to be relatively smooth compared to the example of the Au₁₄₄(SR)₆₀ cluster. The Au₁₄₆(SR)₅₇ does not derive its stability from SAC character; it cannot be considered as a superatom complex.

MicroED Structure of Au146(p-MBA)57 at Subatomic Resolution Reveals a Twinned FCC Cluster

Solving the atomic structure of metallic clusters is fundamental to understanding their optical, electronic, and chemical properties. Herein we present the structure of the largest aqueous gold cluster, Au146(p-MBA)57 (p-MBA: para-mercaptobenzoic acid), solved by electron micro-diffraction (MicroED) to subatomic resolution (0.85 Å) and by X-ray diffraction at atomic resolution (1.3 Å). The 146 gold atoms may be decomposed into two constituent sets consisting of 119 core and 27 peripheral atoms. The core atoms are organized in a twinned FCC structure, whereas the surface gold atoms follow a C2 rotational symmetry about an axis bisecting the twinning plane. The protective layer of 57 p-MBAs fully encloses the cluster and comprises bridging, monomeric, and dimeric staple motifs. Au146(p-MBA)57 is the largest cluster observed exhibiting a bulk-like FCC structure as well as the smallest gold particle exhibiting a stacking fault.