Emerging fungal pathogens represent a concerning threat to both global health and food security. In this study, we aimed to address our rising vulnerability to fungal pathogens through the development of the Fung-AI pipeline: an AI/ML-driven approach for antifungal discovery. A generative adversarial network (GAN) was trained to generate novel candidate antifungal peptide sequences. Next, in silico antifungal and hemolytic classifiers were built to further prioritize AI-generated peptides for experimental validation. From a pool of ~10,000 candidates, thirteen peptides were selected for testing over two-stages of experimentation. Five peptides were found to display mild antifungal activity against the wheat pathogen, Fusarium graminearum, with minimal inhibitory concentrations (MICs) ranging from 250 µg/mL to 500 µg/mL. Four of the five peptides also showed activity against the human pathogen, Candida albicans (MIC: 500 µg/mL). Two of our AI-generated antifungal peptides additionally demonstrated low cytotoxicity in HepG2 human liver carcinoma cells (LC50 > 704.2 µg/mL) indicating that they may be useful as scaffolds for future optimization for therapeutic applications. None of our peptides were found to considerably inhibit the emerging pathogen C. auris, suggesting the need for pathogen-specific down-selection of candidate peptides. Overall, we present a proof-of-principle, generative-AI-based approach for the rapid design of de novo antifungal peptides.

Whole-cell biosensors (WCBs) capable of sensitively detecting trace amounts of analytes hold great potential for in situ detection of pollutants, toxins, or synthetic products. As the terminal signal actuator, the reporter gene directly influences the ultimate sensitivity of WCBs. Although fluorescent proteins (FPs) have been widely used as reporters, their reporting sensitivity is generally lower than that of enzymatic reporters, which often limits the sensitivity and response speed of FP-based sensors in practical applications. Here, we developed an ultrasensitive FP reporter via a noninvasive N-terminal peptide fusion strategy. By adding an N-terminal decapeptide obtained from a high-throughput screening, we constructed an NGFP4 variant that retains the inherent advantages of sfGFP while exhibiting superior reporter gene characteristics, such as rapid expression and robust intracellular stability. These properties enhanced the single-cell fluorescence intensity of NGFP4 by 6.4- to 28-fold in four typical microbial hosts, including E. coli (28-fold), Bacillus subtilis (15.5-fold), Pichia pastoris (9.1-fold), and Saccharomyces cerevisiae (6.4-fold). When applied to WCBs, the NGFP4 reporter greatly shortened the detection time to 1 h for salicylic acid (LOD of 0.36 μM) and 2-chlorobiphenyl (LOD of 18.2 μM), representing the fastest detection time for such sensors. Therefore, our work provides a cross-species compatible FP reporter that enables sensitive detection of microbial cell-based biosensors and other bioanalytical systems, facilitating their field deployment with minimal genetic manipulation and shorter detection time.

Bacteria play fundamental roles in ecosystems, human health, and biotechnology. Although bacterial genome sequencing data have accumulated rapidly over the past decade, the metabolic and ecological functions of most sequenced bacteria remain poorly understood, apart from a few well-studied taxa and traits. Establishing a general framework that comprehensively captures the relationship between bacterial genomes and the diverse biological functions they encode remains a major challenge, as this task requires embedding individual genes within their broader genomic context and modeling the combined effects of gene interactions across complex biological pathways and networks. The difficulty is further compounded by the limited functional annotations available for most bacterial genomes. Here, we introduce BacPT, a bacterial proteome foundation model trained on tens of thousands of complete genomes spanning diverse taxa. BacPT captures both local and genome-wide information, enabling the generation of contextualized gene embeddings and functionally rich representations at the whole organism level. We demonstrate the utility of BacPT across diverse prediction tasks spanning multiple biological scales. BacPT embeddings improve the prediction of enzyme activities, biosynthetic gene clusters BGC, metabolic traits, and ecological interaction outcomes. Our results highlight that unsupervised deep learning applied at the scale of entire proteomes provides a powerful approach for characterizing gene interactions and mapping functional landscapes for bacteria.

Advanced biological imaging analysis platforms such as Activity Quantification and Analysis (AQuA2) enable accurate spatiotemporal activity analysis across diverse cell populations within many species. These tools are increasingly important for investigating cellular signaling dynamics and behavior. However, despite advances in the accuracy and species capability of AQuA2, it remains computationally demanding for analysis of long time-series datasets and requires all users to maintain a MATLAB license, which may limit accessibility and large-scale deployment. To address these limitations, we have designed and made available AQuA2-Cloud, a portable software stack and web platform developed as an improvement and further evolution of AQuA2. This container-deployable system permits multi-user cloud-based high accuracy activity quantification with intuitive workflows, export of analysis data and project files, and comparable processing times. The platform offers integrated features such as in-browser analysis control interfaces, asynchronous program state control, multiple users and user management, support for unreliable connections, file uploading and downloading via web browsers and File Transfer Protocol, and centralized organization of analysis output. AQuA2-Cloud constitutes a cloud-native solution for laboratories or research groups seeking to centralize analysis of spatiotemporal biological imaging datasets while reducing software installation and licensing barriers for end users. The platform enables researchers with minimal technical expertise to perform advanced bioimaging analysis through standard web browsers while maintaining the analytical capabilities of AQuA2. AQuA2-Cloud source code, deployment procedures, and documentation are freely available at (https://github.com/yu-lab-vt/AQuA2-Cloud).

The precise mapping between chemical transformations and enzymatic catalysts underpins the complexity of metabolic networks. Conventional discovery methods, tethered to sequence homology or structural alignment, are inherently blind to new reactions. Here we present VenusRXN, a multimodal deep learning framework that shatters this limitation by enabling reaction-conditioned enzyme discovery. By seamlessly unifying a pre-trained reaction encoder with a protein language model, VenusRXN achieves a fine-grained, high-dimensional alignment of chemical and biological representations. On benchmarks to discover enzymes which catalyze reactions not seen in the training dataset, it surpasses state-of-the-art baselines with a top-20 retrieval hit rate of 76.5%. Most critically, we demonstrate VenusRXN's capabilities in a zero-shot discovery. As verified by the wet-lab experiments, it successfully identified enzymes to catalyze the chemical reactions never reported, including the one to catalyze the synthesis route for a type 2 diabetes drug intermediate using a non-natural substrate. With surprising precision, the model pinpointed active candidates within the top 10 sequences directly from a global search space of over 300 million proteins, which can hardly be achieved by structure-based enzyme discovery algorithm. This work signals a definitive paradigm shift, establishing the chemical reaction itself, rather than homology, as the primary functional descriptor for the de novo discovery of biocatalysts.

Bacteriophages are ubiquitous in nature, but relatively few have been isolated and characterized compared to the number of bacterial strains. Phage biotechnology applications benefit from a diverse library of isolated phages to kill or transfer genetic material to a bacterium of interest. However, scaling phage discovery for diverse bacterial hosts can be time consuming and costly. We developed an approach to capture novel phages for multiple bacteria strains in parallel from an environmental sample using commercially available 0.2-micron filter plates. Using this High-throughput Phage Isolation Platform (HtPIP), we isolated twelve novel phages spanning nine diverse bacterial host genera. Eleven of the isolated phages define new phage species with nine also defining new genera. We show the HtPIP can discover both DNA and RNA phages; including a Tectiviridae infecting Pseudomonas putida mt-2 and a Leviviricetes infecting a Microbacterium isolate, which represents the first cultured RNA phage infecting a host outside of proteobacteria. Using a metagenomic approach, we demonstrate that the HtPIP captures a higher proportion of novel phages compared to traditional low-throughput methods.

Saccharides, a class of essential organic compounds, are ubiquitously found in nature and play a critical role in vital biological processes. They serve as the primary energy source for all living organisms, supporting life functions. In recent years, the synthesis of saccharides via synthetic biology has gained significant attention, with plant systems emerging as a promising alternative to traditional microbial hosts. Plant chassis offer a unique platform by combining photosynthetic carbon fixation with native saccharide biosynthesis and metabolism, enabling sustainable and potentially large-scale saccharide production. This review highlights the advancements in key technologies for saccharide production using plant chassis, summarizes the current research status of plant-based saccharide production across various structural forms, and discusses the technical challenges and strategies for system optimization. The aim is to provide valuable insights for the development of synthetic biology and uncover the commercial potential of plant chassis in saccharide biosynthesis.

Predicting gene function is a pivotal and challenging step in genomic and metagenomic data analysis. Current automatic annotation tools typically rely on the single most similar sequence from the query database. The sparsity of data per annotation makes it challenging to confidently assign gene function for underrepresented genes. Here, we present a contrastive learning framework for functional annotation. FAMUS (Functional Annotation Method Using Supervised contrastive learning) compares query sequences to profile Hidden Markov Model databases and transforms the similarity scores into a condensed vector space that minimizes the distance of proteins from the same family. The similarity scores of a query to all profiles are used for its representation instead of considering only the top-ranking hit. In a protein family assignment task, FAMUS outperformed KEGG's native KofamScan for KEGG Orthology annotation and InterPro's InterProScan for PANTHER family annotation. We thus created four protein annotation models using protein families from the KEGG Orthology, InterPro family, OrthoDB, and EggNOG databases. All four models are available as a conda package and via our user-friendly web server, allowing users to annotate large-scale datasets. FAMUS is the first comprehensive and modular annotation framework based on contrastive learning. It supports both pre-defined and user-specific databases for tailored annotation, and can be easily integrated into any genomic and metagenomic analysis pipeline to facilitate accurate, large-scale functional annotation.

Given that most microbes experience spatially structured environments, examining how such environments affect microbial growth and functions is paramount. Previous studies have shown that a spatially structured environment can impact microbial growth and interactions, and that microbial growth can create or magnify spatial structure. Here, we review some of these instances of past studies to develop a consistent framework that highlights the interplay between microbial interactions, spatial structure of the environment and spatial organization of microbes. We re-examine the level, degree and scale of spatial structure with regard to the phenomena and biological processes of interest. We then discuss how mathematical models can reveal the contribution of the spatial structure to community assembly and coexistence. Lastly, we offer an outlook on important steps for the progress of this field.

Mechanisms by which macrophages deploy antibacterial zinc toxicity are poorly understood. To gain insight into this antimicrobial pathway, we developed bacterial reporter–paired single-cell RNA sequencing of human monocyte-derived macrophages (HMDM) infected with an E. coli zinc-stress reporter strain. We identified HMDM subpopulations harboring zinc-stressed E. coli and corresponding mammalian genes predicted to be associated with either zinc toxicity or survival of zinc-stressed bacteria. Consistent with the latter, SLC30A4 that encodes zinc exporter ZNT4 was enriched in one subpopulation of HMDM containing zinc-stressed E. coli and its overexpression in human macrophages increased intracellular E. coli survival. At a population level, SLC30A4 expression was rapidly downregulated in human macrophages responding to E. coli and its ectopic expression in macrophages attenuated zinc starvation of intracellular E. coli. This is consistent with a model in which macrophages switch off SLC30A4 to engage zinc starvation, while also deploying zinc toxicity against bacteria adapting to a low-zinc environment. Consistent with this, intramacrophage E. coli rapidly upregulated znuA mRNA that is induced during zinc limitation, with zntA mRNA that is induced during zinc stress peaking later. Moreover, E. coli cultured under conditions of zinc limitation displayed greatly enhanced zinc sensitivity. Susceptibility of zinc-sensitive E. coli to killing by macrophages was also attenuated when zinc uptake by E. coli was inactivated, confirming the coordinated actions of zinc starvation and zinc toxicity in macrophage antibacterial responses. Strategies that enhance zinc starvation of intracellular bacteria could be exploited in the design of host-directed therapeutics that amplify macrophage-mediated antibacterial zinc toxicity.

The development of long-read sequencing technologies has enabled the analysis of extended nucleic acid sequences. These methods have proven their strength through their capacity to generate long reads, facilitating the analysis of complex genomic regions and rearrangements. Oxford Nanopore Technologies (ONT) offers a rapid and portable system that brings sequencing to the field. Although this is a great advantage for clinical settings, applications of long-read sequencing in this context have been limited by the high error rates reported for these methods. Here, we report an adaptation of an amplicon sequencing approach combined with unique molecular identifiers. We applied this method to whole-genome sequencing using mock community samples and human blood cultures spiked with common bloodstream infection pathogens. Our results showed a total error rate of <0.1% with V9 chemistry, which was further reduced by <0.05% when using the V14 chemistry. Our results also highlight the improvements of the V14 chemistry on the standard ONT ligation protocol and the importance of the basecalling tool for sequencing accuracy.