The Teleost fish skin microbiome is vital to fish in defending environmental pathogens. Our lab previously found Aeromonas hydrophila (AH) challenge on teleost can induce dysbiotic skin microbiome, leading to skin inflammation. Nevertheless, how the skin-microbiome interaction progresses over time after AH is removed from the environment is not yet understood. This study aims to address this progress under the consideration of skin-microbiome interaction. Here, we conducted a bath challenge of AH in striped catfish for 5 days, followed by AH removal to generate a pathogen-free condition for another 14 days. Fish health status, skin immune responses, AH load in skin mucus, and skin microbiome composition were evaluated by health parameter recording, qPCR, plating, and 16s rRNA gene sequencing respectively before and after challenge, as well as AH removal. We found fish fins reddened and the skin inflammatory cytokine IL-1β increased after challenge but then both significantly decreased after AH removal. Coincidently, AH load in mucus increased after challenge and then decreased significantly after AH removal. The skin microbiome composition, though changed significantly after challenge as previously reported, did not alter significantly after AH removal. Yet, commensal species, such as Limnohabitans, significantly increased in abundance accordingly. In conclusion, these data imply that the host regulation in skin-microbiome interaction may be an essential factor in skin recovery from AH challenge and resulting in skin microbiome change. Taken together, our study may shed some lights in potential developments in post-infection interventions in aquaculture.

Understanding genes and their underlying mechanisms is critical in deciphering how antimicrobial-resistant (AMR) bacteria withstand detrimental effects of antibiotic drugs. At the same time the genes related to AMR phenotypes may also serve as biomarkers for predicting whether a microbial strain is resistant to certain antibiotic drugs. We developed a Cross-Validated Feature Selection (CVFS) approach for robustly selecting the most parsimonious gene set for predicting AMR activities from bacterial pan-genomes. The core idea behind the CVFS approach is interrogating genes (or features in terms of machine learning) among non-overlapping sub-parts of the bacterial gene datasets to ensure the representativeness of the genes. By randomly splitting the data into disjoint sub-parts, conducting machine learning feature selection within each sub-part, and intersecting the features shared by all sub-parts, the CVFS approach is able to achieve the goal of extracting the most representative genes for yielding satisfactory AMR activity prediction accuracy. By testing this idea on bacterial pan-genome datasets, we showed that this approach was able to extract the most succinct feature sets that predicted AMR activities very well, indicating the potential of these genes as AMR biomarker genes. The functional analysis demonstrated that the CVFS approach was able to extract both known AMR genes and novel ones, demonstrating the capabilities of the algorithm in selecting relevant features and highlighting the potential of the novel genes in expanding the antimicrobial resistance databases.

Cyanobacteria are a diverse group of phototrophic microbes with ecological importance. Moreover, these organisms can be harnessed for bioremediation, biofuel production, and production of high-value secondary metabolites. For industrial applications, thermophilic cyanobacteria are particularly valuable as they are a source of enzymes with thermostability properties. To better characterize the diversity of thermophilic cyanobacteria in Taiwan, we conducted environmental sampling across 12 non-acid hot springs. Based on culture-independent survey of 16S ribosomal RNA (rRNA) gene sequences, Cyanobacteria is the most abundant phylum in the microbial mats collected. For culture-dependent survey, Thermosynechococcus is the most dominant genus. Among the 108 newly isolated strains, we selected 27 for whole genome sequencing and found that these strains all belong to a novel species. Gene content analysis indicated a closed pan-genome, suggesting that our sampling likely encompassed most of the genomic diversity of this species. For within-species comparisons, the strains were classified into two distinct groups without obvious geographical separation. In these two groups, 47 and 96 loci were inferred as having experienced selective sweep, respectively. Future characterization of these loci may provide novel insights into the selective forces that drive their genetic divergence. In summary, this work provided a population genomics perspective of the hot spring cyanobacteria in Taiwan, which can serve as a foundation for future development and utilization of biological resources.

Over the past few decades, the abuse of antimicrobial drugs has caused the emergence of multidrug-resistant (MDR) bacteria. However, a well-developed antimicrobial strategy to target MDR bacteria remains explicitly exclusive. In our previous study, we proposed a novel strategy to eliminate target bacteria using programmable inhibitor cells (PICs) that displayed cell adhesion molecules (CAMs) recognizing specific membrane proteins. While eradicating target bacteria by PICs is applicable, it is challenging to identify specific CAMs that recognize desired surface protein antigens on the target cells. In this study, we introduce a whole-cell screening platform to identify CAMs within a synthetic nanobody library.  The platform exploited the potency of conjugation machinery with a positive selection mechanism. The presence of CAMs and their target antigens promotes conjugation even in the liquid environment when cell-cell contacts are transient. This selection results in enrichment for bacteria displaying cognate CAMs. As a result, we identified CAMs targeting surface antigens TraN, OmpA, and OmpC in their native conformations. PICs displayed these CAMs precisely deplete targeted bacteria within microbial communities. In conclusion, this platform uncovers the potential of tailored bacterial interactions and opens up a new era of synthetic applications in basic research, biotechnology, and clinical settings.

Biovar 1 agrobacteria are soil-borne phytopathogens that can cause crown gall disease and are utilized in plant transformation. One unusual trait of these bacteria is their multipartite genomes, with each cell containing one circular chromosome and one linear chromid as essential replicons. To better understand the evolution of such multipartite genomes, we analyzed 26 complete genome assemblies from 9 Agrobacterium species. Based on gene content analysis, the chromosomes are more conserved in harboring 1,711 core genes, 873 species-specific genes, and 6,267 accessory genes. In comparison, the chromids harbor 357 core genes, 1,691 species-specific genes, and 7,085 accessory genes. Unexpectedly, the core gene phylogenies of these two replicons are incongruent in the placement of Agrobacterium tomkonis (genomospecies 3). Moreover, the species boundaries are better delineated based on the chromid gene content, rather than that of the chromosome. Based on the nonsynonymous to synonymous substitution rate ratio (dN/dS) of core genes, one ~284-kb region of the chromosome was found to have experienced relaxed purifying selection, while no such evolutionary hotspot was identified in the chromid. Taken together, our findings indicated that the two essential replicons of biovar 1 agrobacteria are different in their evolutionary history and patterns of molecular evolution, which provided novel understanding of these importance microbes. The roles of those differences in adaptation, as well as if there are general principles for the evolution of bacteria with multipartite genomes, require further investigation.

The climate change brings out several environmental issues, water scarcity is one of a severe problem. The urban waste water reusing might be a new source of it, especially in agriculture, which is the most water-consuming industry. Previous study already indicated that application of reused wastewater increased the concentration of nutrients in the soil, however the effect on soil microbial community and the risks are still unclear, so we used pot experiment to reveal the effect of irrigating by reused water. We collected two different sites of reusing water, GT and LY, compared with tap water irrigation (CT), as well as the plant model is Oryza sativa ssp. japonica. Results showed that irrigation with GT and LY increase the relative abundance of plant growth promoting bacteria and the pollutant-degrading bacteria. The alpha-diversity revealed that both the richness and diversity index were higher in LY than in the irrigation with GT and CT, and beta-diversity showed by the principal co-ordinates plot based on Bray-Curtis distance indicated that reusing water significantly changed the microbial community. Furthermore, the microbiomes clustered into 3 group by different irrigation sources, each treatment involved unique taxa, particularly the N-cycling relative microbes appeared in the reusing wastewater treatment (LY and GT). In the future, we will design field experiments to investigate the effect in a bigger scale, and hope that we will be able to utilize a high quality and healthy new water source.

Microbial extracellular polymeric substances (exopolymers) are a complex mixture of biopolymers consisting of polysaccharides, proteins, nucleic acids, and other metabolites, important for many biological activities including bacteria-bacteria interaction. In our preliminary data, Pseudomonas taiwanensis (PT), a commensal bacterium that resides in the striped catfish skin mucus, showed inhibition activity toward a common striped catfish pathogen Aeromonas hydrophila (AH) in vitro. However, their interaction in the striped catfish skin mucus during AH pathogenesis is still unknown. Here, we employ PT exopolymers to investigate the role of PT on AH for bacteria-bacteria interaction during AH pathogenesis in the striped catfish skin mucus. PT exopolysaccharide (EPS) and cell-free supernatant (CFS) were evaluated by incubating AH in vitro with or without EPS or CFS, and their effect on AH were examined using fluorescence microscope and live-cell imaging. Under EPS and CFS treatment, AH adherence was significantly reduced compared to non-treated control. Moreover, AH showed phenotypic change in CFS treatment by forming aggregation. In contrast, AH motility did not significantly change compared to non-treated control. We then measured the expression of toxin-associated genes for AH pathogenesis by qPCR. We observed that EPS and CFS treatment also influenced the expressions of aerolysin (aerA), hemolysin (ahh1), and elastase (ela) genes. In conclusion, these data indicate that PT interaction with AH in striped catfish skin mucus could potentially interfere the AH pathogenesis. Thus, this study provides information on the potential use of PT as a biological agent for disease prevention in aquaculture.

Fish skin mucus microbiome is important in preventing the aquatic pathogens. Previous studies have reported that both environment and host can influence the fish skin mucus microbiome dynamics. However, how environment or host alone interacts with skin mucus microbiome is unknown. To address this issue, we seek to develop the in vitro fish skin mucus microbiome to test environment or host factor separately. Here, we selected seven media including brain heart infusion (BHI) broth with or without additives (galactose, inorganic salts, Tween80), Leibovitz’s L-15 (L15), and Dulbecco's Modified Eagle Medium (DMEM) as culture medium for skin mucus. Fish skin mucus was collected and inoculated into each medium for culturing 48 h with and without passaging at 24 h. Samples were collected at 24 and 48 h followed by 16S rRNA gene sequencing for microbiome community analysis. We found that skin mucus microbiome culture in BHI with inorganic salts can maintain 59.5% amplicon sequence variant (ASV) levels while culture in other medium can only maintain ASV levels up to 26.1% - 52.3%. Furthermore, we found that the 48 h culture without passaging exhibits an additional 19% increase in ASV levels compare to 24 h in BHI with inorganic salts while less increase or even decrease in other media. In conclusion, culturing skin mucus microbiome in BHI with inorganic salts for 48 h without passaging can maintain the highest microbiome abundance thereby serving as the suitable medium for in vitro model. Yet, the evenness of cultured mucus microbiome is reduced with the loss of major environmental-oriented bacterial genera compared to that of the fresh mucus. Thus, these data suggest that adding environmental or the corresponding host factors in the medium should be considered in future work. This study provides preliminary information in the development of in vitro fish skin microbiome model.

The marine microbial food web mediates the transport of deep-sea active carbon through its capability of diel vertical migration (DVM). However, limited studies have explored the vertical distributions and migrations of marine plankton, with few considerations for temporal variability across vertical water layers. This project sampled and profiled the vertical distribution of plankton communities at multiple time points over one day. Prokaryotic or eukaryotic plankton were successfully enriched by size screening. Prokaryotic plankton communities at the surface (<10 M) and deep (>80 M) water layers showed relatively little diel variation, whereas the midwater layers (20-75 M) showed large changes in community composition over one day. According to the overall temporal and vertical patterns, we can further define plankton taxa into three main types: surface-fixed, deep-fixed and diel-moving. This project systemically revealed species-specific plankton DVM and reconstruct ecological interactions among various plankton groups in marine ecosystems.

珊瑚共生體是由珊瑚和許多微生物組成的動態系統，系統中的細菌和珊瑚存在許多涉及珊瑚健康的互動，包含珊瑚的代謝過程、初級代謝物和次級代謝物的合成。初級代謝物是指直接涉及到正常生長、發育與生殖的代謝產物，次級代謝物則是由初級代謝物衍生化的化學物質。目前已知，在珊瑚適應環境變化的壓力下，這些代謝物扮演著重要角色。近年來，人類活動、全球氣候變遷和溫室效應的加劇，珊瑚礁生態系逐漸退化，如何以復育珊瑚一途來恢復珊瑚礁生態系成了重要課題，而異地珊瑚養殖是其中一種方法，然而異地養殖的方式是否會影響珊瑚共生體與其代謝產物的相關研究並不多。此外，由於八放珊瑚所含的次級代謝物具有比石珊瑚多元，因此，深入了解八放珊瑚共生菌與宿主生理及在原生地與異地之次級代謝物產生的差異，也將有助於珊瑚復育與次級代謝物資源的開發，也有助於之後珊瑚模式物種的選擇與相關研究。因此，在本研究中，我們調查了龜山島的兩種八放珊瑚Lobophytum sp.與Sclerophytum sp. 在人工養殖和龜山島原生環境生長之珊瑚細菌菌相和初級代謝物氨基酸和脂肪酸組成之間的差異。我們預期隨著養殖時間達到兩個月後，珊瑚菌相組成和初級代謝物組成會逐漸趨於穩定，並且與原生環境中珊瑚的差異趨於相似。此外，兩種八放珊瑚的趨勢可能會因物種差異而有所不同。本研究藉由比較得知人工環境下八放珊瑚生殖生理變化與微生物族群變動所需時間，對八放珊瑚初級代謝物的生成與微生物的關係奠下基石，相信此成果將對於軟珊瑚天然物生合成的研究將有所助益。

Methane is a powerful greenhouse gas with 28-34 times greater global warming potential than carbon dioxide. Recent studies showed that marine microbial communities can consume dissolved methane before it can escape into the atmosphere. In shallow layer marine sediment, aerobic methanotrophs which can consume methane to methanol through enzyme pMMO were found through metagenomics, Still, studies focused on methane consuming potential of coral reef sediment were absent. So, our study will focus on the aerobic methanotrophs composition in coral reef sediments and their methane oxidizing potential. Turtle Island, our study site is a special geological area that has a shallow hydrothermal vent at the front of Turtle Island (Head). At the same time, the coral reef was located behind Turtle Island (Tail). This study site provides two different ecosystems in the same area and we can compare their aerobic methanotroph composition and methane-oxidizing potential at the same time. Here we show that sediments of two ecosystems have different methane oxidizing potential. At the Tail site (coral reef), the methane-oxidizing potential is higher than at the Head site (hydrothermal vent). In contrast, the South side, located between the Tail and Head sites, has different methane oxidizing potential in two seasons, showing that the South site was a transition site that will be affected by both the Tail and Head sites. In sum, the coral reef ecosystem has methane oxidizing potential, compared to the shallow hydrothermal vent in Turtle Island.

The composition of symbiotic algae and microbial communities is closely associated with the health and resilience of coral holobionts. Compared with stony corals, octocorals have better environmental resistance, however, the relationship between members of octocoral holobiont is not understood. Through a three-year observational study, we found significant bleaching events after summer, followed by recovery in winter among the octocoral Litophyton in the Bitou Cape of northeastern Taiwan. This prompted our interest in understanding the relationship between octocoral bleaching and the composition of its symbiotic organisms. 

In this research, we used next-generation sequencing to decipher the ITS2 and 16S rRNA V6-V8 region of the coral holobiont. So far, we have completed two year analysis of the octocoral microbiome. We found that the relative abundance of Endozoicomonas increased during coral bleaching, and the proportion of Terasakiellaceae decreased. Conversely, when the coral is healthy, the trend is reversed. In beta diversity analysis, there is a significant seasonal difference in Litophyton microbial composition, and the microbial composition became increasingly dissimilar as the seasons transition from spring to winter. Investigating variations in the composition of the holobiont enhances our comprehension of the interplay among its members. Simultaneously, long-term observations can help us gain a deeper understanding of the adaptability and resilience of octocorals to environmental changes brought about by climate change.

在自然環境中，微生物與植物的生長息息相關，番茄微生物菌相容易受到水、空氣和土壤等環境因子的影響，針對番茄根部和葉片探討微生物菌相的改變是否會影響番茄的生長狀態，因此我們衍生出了兩個研究方向，(1) 使用番茄致病菌感染宿主植物，並使用益生菌作為預防和改善，分析葉片微生物菌相在受到病原菌和益生菌的影響之下的動態變化，並觀察它與生長狀態的關聯性 (2) 進一步探討植物根部益生菌與番茄之間的交互作用，特別關注番茄可能透過調控植物的醣傳運基因為根部微生物群提供碳源的機制，進而影響根部微生物群的組成與多樣性；我們的研究聚焦於根部微生物群的相互作用，這可能對番茄生長狀態和健康產生深遠影響，這些發現將有助於拓展對番茄微生態系統的全面認識。

In the natural environment, the growth of plants is intricately linked to the presence of microorganisms. The tomato microbiota is susceptible to environmental factors such as water, air, and soil. We investigate the impact of changes in microbial communities on tomato growth, focusing on both roots and leaves. Two research directions emerge: (1) employing pathogenic bacteria to infect host plants and utilizing probiotics for prevention and improvement, analyzing dynamic changes in leaf microbial communities under the influence of pathogens and probiotics, correlating with growth status; (2) delving into the interaction between plant root probiotics and tomatoes, particularly emphasizing the potential mechanism where tomatoes regulate plant sugar transport genes to provide carbon sources for root microbial communities, thereby influencing the composition and diversity of root microbial communities. Our research concentrates on the interplay of root microbial communities, with potential profound effects on tomato growth and health. These findings contribute to a comprehensive understanding of the tomato microbiota ecosystem.

The persistent overuse of antibiotics, especially for minor infections, is diminishing their effectiveness in treating serious infections. Consequently, individuals are succumbing to common infections that were once treatable. The issue of antibiotic resistance (AMR) has emerged as a significant challenge in modern medicine. Patients must resort to higher drug doses to attain efficacy, thereby becoming ensnared in a cycle of escalating doses. To prevent this scenario, there is a need to address the problem of antimicrobial resistances. The antibiotic resistance of microbial pathogens is primarily associated with their genetic makeup. One of the most prevalent methods for identifying AMR pathogens involves searching for known AMR genes through comparisons with databases, allowing for the inference of resistance. However, in previous works we have shown that genes selected by machine learning feature selection approaches achieve better prediction accuracy than known AMR gene sets, highlighting the effectiveness and usefulness of machine learning-selected genes.

In this study, we extended the feature selection idea by designing a novel feature selection methodology, in which a feature selection algorithm (eXtreme Graduent Boosting; XGBoost) is employed as the foundational architectural operation while the prediction accuracy of each gene is inspected to identify the most effective and parsimonious combination of gene sets in predicting resistances. By testing this approach on bacterial gene sets we found that the devised methodology selected a much smaller gene set than others (p-value < 0.01) while achieving very similar prediction performances, suggesting that the selected gene set is more effective in predicting AMR status and underscoring the feasibility of the proposed method for more accurate identification of important genes related to antibiotic resistance.

The type VI secretion system (T6SS) is a widespread contractile nanoweapon in Gram-negative bacteria to gain niche advantages by translocating effectors into recipient cells. Agrobacterium tumefaciens strain C58 harbors two T6SS DNase effectors (Tde1 and Tde2) as the major antibacterial toxins to kill the recipient cells. Interestingly, some factors produced in the recipient cells can enhance the killing outcome, therefore named recipient susceptibility (RS) factors. Previous screening of Escherichia coli BW25113 Keio library identified ClpA ATPase and ClpP protease functioning to form a proteolysis complex are required to enhance C58 T6SS-dependent killing. In this study, we found that the substrate adaptor, ClpS, is also an RS factor and the ClpAPS is enhancing Tde2-mediated T6SS killing. Importantly, clpP or clpS mutants of phytopathogen Dickeya dadantii and A. tumefaciens C58 also confer resistance to Tde2-mediated T6SS killing, suggesting ClpAPS protease complex is a general RS factor. One plausible hypothesis is that the ClpAPS may regulate Tde2 killing activity by degrading its substrates that inhibit Tde2 DNase activity via direct binding. Current results showed that Tde2-mediated growth inhibition is recovered in the absence of ClpA, ClpP, or ClpS, and ClpA could be co-purified with Tde2 when ectopic expression in E. coli. Through Tde2 immunoprecipitation coupled with mass spectrometry analysis, six known ClpAPS substrates were identified as Tde2-interacting protein candidates. The role of the six ClpAPS substrates involved in Tde2-mediated killing is under investigation.