Versatile Framework for Drug-Target Interaction Prediction by Considering Domain-Specific Features.

Predicting drug-target interactions (DTIs) is one of the crucial tasks in drug discovery, but traditional wet-lab experiments are costly and time-consuming. Recently, deep learning has emerged as a promising tool for accelerating DTI prediction due to its powerful performance. However, the models trained on limited known DTI data struggle to generalize effectively to novel drug-target pairs. In this work, we propose a strategy to train an ensemble of models by capturing both domain-generic and domain-specific features (E-DIS) to learn diverse domain features and adapt them to out-of-distribution data. Multiple experts were trained on different domains to capture and align domain-specific information from various distributions without accessing any data from unseen domains. E-DIS provides a comprehensive representation of proteins and ligands by capturing diverse features. Experimental results on four benchmark data sets in both in-domain and cross-domain settings demonstrated that E-DIS significantly improved model performance and domain generalization compared to existing methods. Our approach presents a significant advancement in DTI prediction by combining domain-generic and domain-specific features, enhancing the generalization ability of the DTI prediction model.

Unsupervisedly Prompting AlphaFold2 for Accurate Few-Shot Protein Structure Prediction.

Data-driven predictive methods that can efficiently and accurately transform protein sequences into biologically active structures are highly valuable for scientific research and medical development. Determining an accurate folding landscape using coevolutionary information is fundamental to the success of modern protein structure prediction methods. As the state of the art, AlphaFold2 has dramatically raised the accuracy without performing explicit coevolutionary analysis. Nevertheless, its performance still shows strong dependence on available sequence homologues. Based on the interrogation on the cause of such dependence, we presented EvoGen, a meta generative model, to remedy the underperformance of AlphaFold2 for poor MSA targets. By prompting the model with calibrated or virtually generated homologue sequences, EvoGen helps AlphaFold2 fold accurately in the low-data regime and even achieve encouraging performance with single-sequence predictions. Being able to make accurate predictions with few-shot MSA not only generalizes AlphaFold2 better for orphan sequences but also democratizes its use for high-throughput applications. Besides, EvoGen combined with AlphaFold2 yields a probabilistic structure generation method that could explore alternative conformations of protein sequences, and the task-aware differentiable algorithm for sequence generation will benefit other related tasks including protein design.

Compact and ultrastable photonic microwave oscillator.

Frequency comb synthesized microwaves have been so far realized with tabletop systems, operated in well-controlled environments. Here, we demonstrate state-of-the-art ultrastable microwave synthesis with a compact rack-mountable apparatus. We present absolute phase noise characterization of a 12 GHz signal using an ultrastable laser at $\sim{194}\;{\rm THz}$∼194THz and an Er:fiber comb divider, obtaining $ - {83}\;{\rm dBc/Hz}$-83dBc/Hz at 1 Hz and $ \lt - {166}\;{\rm dBc/Hz}$<-166dBc/Hz for offsets greater than 5 kHz. Employing semiconductor coating mirrors for the same type of transportable optical frequency reference, we show that $ - {105}\;{\rm dBc/Hz}$-105dBc/Hz at 1 Hz is supported by demonstrating a residual noise limit of division and detection process of $ - {115}\;{\rm dBc/Hz}$-115dBc/Hz at 1 Hz. This level of fidelity paves the way for the deployment of ultrastable photonic microwave oscillators and for operating transportable optical clocks.

Detection of Exosomal PD-L1 RNA in Saliva of Patients With Periodontitis.

Periodontitis is the most prevalent inflammatory disease of the periodontium, and is related to oral and systemic health. Exosomes are emerging as non-invasive biomarker for liquid biopsy. We here evaluated the levels of programmed death-ligand 1 (PD-L1) mRNA in salivary exosomes from patients with periodontitis and non-periodontitis controls. The purposes of this study were to establish a procedure for isolation and detection of mRNA in exosomes from saliva of periodontitis patients, to characterize the level of salivary exosomal PD-L1, and to illustrate its clinical relevance. Bioinformatics analysis suggested that periodontitis was associated with an inflammation gene expression signature, that PD-L1 expression positively correlated with inflammation in periodontitis based on gene set enrichment analysis (GSEA) and that PD-L1 expression was remarkably elevated in periodontitis patients versus control subjects. Exosomal RNAs were successfully isolated from saliva of 61 patients and 30 controls and were subjected to qRT-PCR. Levels of PD-L1 mRNA in salivary exosomes were higher in periodontitis patients than controls (P < 0.01). Salivary exosomal PD-L1 mRNA showed significant difference between the stages of periodontitis. In summary, the protocols for isolating and detecting exosomal RNA from saliva of periodontitis patients were, for the first time, characterized. The current study suggests that assay of exosomes-based PD-L1 mRNA in saliva has potential to distinguish periodontitis from the healthy, and the levels correlate with the severity/stage of periodontitis.

Multivariate multiscale complex network analysis of vertical upward oil-water two-phase flow in a small diameter pipe.

High water cut and low velocity vertical upward oil-water two-phase flow is a typical complex system with the features of multiscale, unstable and non-homogenous. We first measure local flow information by using distributed conductance sensor and then develop a multivariate multiscale complex network (MMCN) to reveal the dispersed oil-in-water local flow behavior. Specifically, we infer complex networks at different scales from multi-channel measurements for three typical vertical oil-in-water flow patterns. Then we characterize the generated multiscale complex networks in terms of network clustering measure. The results suggest that the clustering coefficient entropy from the MMCN not only allows indicating the oil-in-water flow pattern transition but also enables to probe the dynamical flow behavior governing the transitions of vertical oil-water two-phase flow.