The ectomycorrhizal fungus Laccaria bicolor is a key symbiotic mutualist in forest ecosystems, where it enhances nutrient uptake and promotes the growth of host trees. Here, we present genome assemblies of 14 geographically distinct strains and progeny of L. bicolor, providing new insights into the intraspecific genomic diversity. Pangenome analysis revealed substantial variation in assembly size (42-96 Mbp), gene content (16,084-26,800 genes), and single nucleotide polymorphism (SNP) density (0.04-12.08 SNPs/kb). This variation likely reflects genuine biological differences among strains adapted to diverse environmental conditions, although differences in assembly quality and repeat content may also play a role. These genomic resources, comprising draft genome assemblies with comprehensive annotations, will facilitate comparative studies of the genetic diversity and functional traits underlying the ecological success of this model ectomycorrhizal fungus.

Many macrofungi are impractical or impossible to culture. Consequently, DNA for long-read sequencing required for the assembly of high-quality genomes must be isolated from samples taken from the environment. Collection is often in remote locations, limiting the options for stabilising samples to methods that do not require refrigeration. Fungi contain species-specific arrays of metabolites that may complicate purification techniques and call for judgement to be made to apply appropriate modifications to the DNA extraction protocol in specific cases. The protocols and commentary we describe are informed by the preparation of DNA from a range of Australasian ectomycorrhizal and saprotrophic macrofungi. We collect samples into isopropanol at ambient temperature and employ a strategy of chromatin isolation followed by the sequential removal of unwanted molecular components to purify DNA.

The non-linear nature of deep neural networks makes it difficult to interpret the reason behind their output, thus reducing verifiability of the system where these models are applied. Understanding the patterns between activation vectors and predictions could give insight as to erroneous classifications and how to identify them. This paper explains a systematic approach to identifying the clusters with the most misclassifications or false label annotations. For this research, we extracted the activation vectors from a deep learning model, DNABERT, and visualized them using t-SNE to decode the reason behind the results that are produced. We applied K-means in a hierarchical fashion on the activation vectors for a set of training instances. We analyzed cluster mean activation vectors to find any patterns in the errors across K-means clusters. The cluster analysis revealed that the predictions were uniform, or nearly 100 percent the same, in clusters of similar activation vectors. It was found that two clusters containing most of their objects belonging to the same true class tend to be closer together than clusters of opposite classes. The means of objects of the same true label are closer if two clusters have the same predicted labels rather than opposite predicted labels, showing that the activation vectors reflect both predicted and true classes. We did a similar analysis for all 26 organisms in the dataset, showing the Euclidean distance can be used for identifying clusters with many errors. We propose a heuristic to find the clusters with a high number of misclassifications or incorrect label annotations using the vector analysis between clusters. This can aid in identifying misclassifications of DNA sequences or problems with sequence tagging.

Detecting social engineering attempts is crucial for Internet security. Such threats are becoming more frequent and more strategic in their exploitation of human vulnerabilities. Our research focuses on topic modeling of social engineering attacks through the use of conversational scenarios that illustrate various strategies. These dataset scenarios have been manually extracted from a body of text and synthetically augmented to ensure natural dialogue flow. Two methodologies are presented for adaptive utterance- and global-level topic extraction: prompt engineering with OpenAI's GPT-4o-mini, characterized by fewshot learning and chain-of-thought prompting, and efficient fine-tuning utilizing Quantized Low Rank Adaptation (QLoRA) with Mistral's 7B Instruct model. Through experimentation and evaluation, this study aims to determine the effectiveness of these topic modeling approaches in enhancing the vital task of social engineering scenario identification in a real-time framework.

Lorchels, also known as false morels (Gyromitra sensu lato), are iconic due to their brain-shaped mushrooms and production of gyromitrin, a deadly mycotoxin. Molecular phylogenetic studies have hitherto failed to resolve deep-branching relationships in the lorchel family, Discinaceae, hampering our ability to settle longstanding taxonomic debates and to reconstruct the evolution of toxin production. We generated 75 draft genomes from cultures and ascomata (some collected as early as 1960), conducted phylogenomic analyses using 1542 single-copy orthologs to infer the early evolutionary history of lorchels, and identified genomic signatures of trophic mode and mating-type loci to better understand lorchel ecology and reproductive biology. Our phylogenomic tree was supported by high gene tree concordance, facilitating taxonomic revisions in Discinaceae. We recognized 10 genera across two tribes: tribe Discineae (Discina, Maublancomyces, Neogyromitra, Piscidiscina, and Pseudodiscina) and tribe Gyromitreae (Gyromitra, Hydnotrya, Paragyromitra, Pseudorhizina, and Pseudoverpa); Piscidiscina was newly erected and 26 new combinations were formalized. Paradiscina melaleuca and Marcelleina donadinii formed their own family-level clade sister to Morchellaceae, which merits further taxonomic study. Genome size and CAZyme content were consistent with a mycorrhizal lifestyle for the truffle species (Hydnotrya spp.), whereas the other Discinaceae genera possessed genomic properties of a saprotrophic habit. Lorchels were found to be predominantly heterothallic—either MAT1-1 or MAT1-2—but a single occurrence of colocalized mating-type idiomorphs indicative of homothallism was observed in Gyromitra esculenta strain CBS101906 and requires additional confirmation and follow-up study. Lastly, we confirmed that gyromitrin has a phylogenetically discontinuous distribution, having been detected exclusively in two distantly related genera (Gyromitra and Piscidiscina) belonging to separate tribes. Our genomic dataset will facilitate further investigations into the gyromitrin biosynthesis genes and their evolutionary history. With additional sampling of Geomoriaceae and Helvellaceae—two closely related families with no publicly available genomes—these data will enable comprehensive studies on the independent evolution of truffles and ecological diversification in an economically important group of pezizalean fungi.

Infections caused by antimicrobial-resistant Escherichia coli are the leading cause of death attributed to antimicrobial resistance (AMR) worldwide, and the known AMR mechanisms involve a range of functional proteins. Here, we employed a pan-genome wide association study (GWAS) approach on over 1,000 E. coli isolates from sick dogs collected across the US and Canada and identified a strong statistical association (empirical P < 0.01) of AMR, involving a range of antibiotics to a group 1 capsular (CPS) gene cluster. This cluster included genes under relaxed selection pressure, had several loci missing, and had pseudogenes for other key loci. Furthermore, this cluster is widespread in E. coli and Klebsiella clinical isolates across multiple host species. Earlier studies demonstrated that the octameric CPS polysaccharide export protein Wza can transmit macrolide antibiotics into the E. coli periplasm. We suggest that the CPS in question, and its highly divergent Wza, functions as an antibiotic trap, preventing antimicrobial penetration. We also highlight the high diversity of lineages circulating in dogs across all regions studied, the overlap with human lineages, and regional prevalence of resistance to multiple antimicrobial classes.

Large language models (LLMs) have exploded in popularity recently, becoming widely used for various tasks. However, concerns persist regarding potential biases within these models. This study investigated gender bias in the popular LLMs such as GPT-3.5, GPT-4, Gemini, and Meta's Llama2. In this study, we focused on analyzing biases using ambiguous sentences across English, Malayalam, and Tamil languages. We evaluated the LLMs to see if they associate occupations with commonly held gender stereotypes, by using specific professions within our test sentences. We examined differences between the biases in the three languages in LLMs. Our findings suggested that strong gender biases exist in all the LLMs across the three languages. This study can be used to understand the biases in large language models and can be equipped to use AI in a way that mitigates bias.

Deep learning struggles with unsupervised tasks like community detection in networks. This work proposes the Enhanced Community Detection with Structural Information VGAE (VGAE-ECF) method, a method that enhances variational graph autoencoders (VGAEs) for community detection in large networks. It incorporates community structure information and edge weights alongside traditional network data. This combined input leads to improved latent representations for community identification via K-means clustering. We perform experiments and show that our method works better than previous approaches of community-aware VGAEs.

This study aims to simplify Environmental Health and Safety (EHS) by leveraging the power of Large Language Models (LLMs). In this research, we focus on fine-tuning three LLMs — Llama, Mistral, and Falcon — using PEFT techniques such as QLoRA and SFT, to address domain-specific needs such as safety compliance, incident reporting, and knowledge dissemination. Our research methodology involves fine-tuning each LLM model on a custom dataset compiled from various regulatory agencies, supplemented by targeted web scraping and manual collection of questionnaires to capture and enrich the models with the latest regulations and guidelines. This study aims to compare the effectiveness of these fine-tuned models to identify the most effective model and fine-tuning techniques for specific EHS applications. We found that Mistral outperformed Falcon and Llama on ROUGE, BERTscore, and a manual evaluation.

This paper presents a novel approach to enhancing the Llama language model for generating customized resumes tailored to domain-specific job descriptions. Unlike traditional methods that rely heavily on extensive fine-tuning, we implement scalable adapter modules to minimize parameter adjustments. This approach preserves the model's inherent ability to generate coherent and contextually appropriate language across diverse tasks, ensuring that its general linguistic capabilities remain intact. Additionally, we employ a prompting strategy to dynamically create a diverse and comprehensive dataset, ensuring high relevance to various job roles. Using a dataset of 10,000 job descriptions and resumes, our approach resulted in a 20% improvement in BLEU scores and a 70% reduction in perplexity compared to the base model. This combination allows for an efficient and effective fine-tuning process, resulting in superior performance in generating job-specific resume content as evidenced by improved BLEU scores and perplexity metrics. Our approach provides a practical solution for scalable production deployment, while maintaining the model's adaptability and robustness across different job domains.

The Lipomyces clade contains oleaginous yeast species with advantageous metabolic features for biochemical and biofuel production. Limited knowledge about the metabolic networks of the species and limited tools for genetic engineering have led to a relatively small amount of research on the microbes. Here, a genome-scale metabolic model (GSM) of Lipomyces starkeyi NRRL Y-11557 was built using orthologous protein mappings to model yeast species. Phenotypic growth assays were used to validate the GSM (66% accuracy) and indicated that NRRL Y-11557 utilized diverse carbohydrates but had more limited catabolism of organic acids. The final GSM contained 2,193 reactions, 1,909 metabolites, and 996 genes and was thus named iLst996. The model contained 96 of the annotated carbohydrate-active enzymes. iLst996 predicted a flux distribution in line with oleaginous yeast measurements and was utilized to predict theoretical lipid yields. Twenty-five other yeasts in the Lipomyces clade were then genome sequenced and annotated. Sixteen of the Lipomyces species had orthologs for more than 97% of the iLst996 genes, demonstrating the usefulness of iLst996 as a broad GSM for Lipomyces metabolism. Pathways that diverged from iLst996 mainly revolved around alternate carbon metabolism, with ortholog groups excluding NRRL Y-11557 annotated to be involved in transport, glycerolipid, and starch metabolism, among others. Overall, this study provides a useful modeling tool and data for analyzing and understanding Lipomyces species metabolism and will assist further engineering efforts in Lipomyces.

Background: Genome assembly tools are used to reconstruct genomic sequences from raw sequencing data, which are then used for identifying the organisms present in a metagenomic sample. Methodology: More recently, machine learning approaches have been applied to a variety of bioinformatics problems, and in this paper, we explore their use for organism identification. We start by evaluating several commonly used metagenomic assembly tools, including PhyloFlash, MEGAHIT, MetaSPAdes, Kraken2, Mothur, UniCycler, and PathRacer, and compare them against state-of-theart deep learning-based machine learning classification approaches represented by DNABERT and DeLUCS, in the context of two synthetic mock community datasets. Result: Our analysis focuses on determining whether ensembling metagenome assembly tools with machine learning tools have the potential to improve identification performance relative to using the tools individually. Conclusion: We find that this is indeed the case, and analyze the level of effectiveness of potential tool ensembling for organisms with different characteristics (based on factors such as repetitiveness, genome size, and GC content).

Mycena s.s. is a ubiquitous mushroom genus whose members degrade multiple dead plant substrates and opportunistically invade living plant roots. Having sequenced the nuclear genomes of 24 Mycena species, we find them to defy the expected patterns for fungi based on both their traditionally perceived saprotrophic ecology and substrate specializations. Mycena displayed massive genome expansions overall affecting all gene families, driven by novel gene family emergence, gene duplications, enlarged secretomes encoding polysaccharide degradation enzymes, transposable element (TE) proliferation, and horizontal gene transfers. Mainly due to TE proliferation, Arctic Mycena species display genomes of up to 502 Mbp (2–8× the temperate Mycena), the largest among mushroom-forming Agaricomycetes, indicating a possible evolutionary convergence to genomic expansions sometimes seen in Arctic plants.Overall, Mycena show highly unusual, varied mosaic-like genomic structures adaptable to multiple lifestyles, providing genomic illustration for the growing realization that fungal niche adaptations can be far more fluid than traditionally believed.

Thermophily is a trait scattered across the fungal tree of life, with its highest prevalence within three fungal families (Chaetomiaceae, Thermoascaceae, and Trichocomaceae), as well as some members of the phylum Mucoromycota. We examined 37 thermophilic and thermotolerant species and 42 mesophilic species for this study and identified thermophily as the ancestral state of all three prominent families of thermophilic fungi. Thermophilic fungal genomes were found to encode various thermostable enzymes, including carbohydrate-active enzymes such as endoxylanases, which are useful for many industrial applications. At the same time, the overall gene counts, especially in gene families responsible for microbial defense such as secondary metabolism, are reduced in thermophiles compared to mesophiles. We also found a reduction in the core genome size of thermophiles in both the Chaetomiaceae family and the Eurotiomycetes class. The Gene Ontology terms lost in thermophilic fungi include primary metabolism, transporters, UV response, and O-methyltransferases. Comparative genomics analysis also revealed higher GC content in the third base of codons (GC3) and a lower effective number of codons in fungal thermophiles than in both thermotolerant and mesophilic fungi. Furthermore, using the Support Vector Machine classifier, we identified several Pfam domains capable of discriminating between genomes of thermophiles and mesophiles with 94% accuracy. Using AlphaFold2 to predict protein structures of endoxylanases (GH10), we built a similarity network based on the structures. We found that the number of disulfide bonds appears important for protein structure, and the network clusters based on protein structures correlate with the optimal activity temperature. Thus, comparative genomics offers new insights into the biology, adaptation, and evolutionary history of thermophilic fungi while providing a parts list for bioengineering applications.

Background Cost-effective production of biofuels from lignocellulose requires the fermentation of d -xylose. Many yeast species within and closely related to the genera Spathaspora and Scheffersomyces (both of the order Serinales) natively assimilate and ferment xylose. Other species consume xylose inefficiently, leading to extracellular accumulation of xylitol. Xylitol excretion is thought to be due to the different cofactor requirements of the first two steps of xylose metabolism. Xylose reductase (XR) generally uses NADPH to reduce xylose to xylitol, while xylitol dehydrogenase (XDH) generally uses NAD + to oxidize xylitol to xylulose, creating an imbalanced redox pathway. This imbalance is thought to be particularly consequential in hypoxic or anoxic environments. Results We screened the growth of xylose-fermenting yeast species in high and moderate aeration and identified both ethanol producers and xylitol producers. Selected species were further characterized for their XR and XDH cofactor preferences by enzyme assays and gene expression patterns by RNA-Seq. Our data revealed that xylose metabolism is more redox balanced in some species, but it is strongly affected by oxygen levels. Under high aeration, most species switched from ethanol production to xylitol accumulation, despite the availability of ample oxygen to accept electrons from NADH. This switch was followed by decreases in enzyme activity and the expression of genes related to xylose metabolism, suggesting that bottlenecks in xylose fermentation are not always due to cofactor preferences. Finally, we expressed XYL genes from multiple Scheffersomyces species in a strain of Saccharomyces cerevisiae . Recombinant S. cerevisiae expressing XYL1 from Scheffersomyces xylosifermentans , which encodes an XR without a cofactor preference, showed improved anaerobic growth on xylose as the primary carbon source compared to S. cerevisiae strain expressing XYL genes from Scheffersomyces stipitis . Conclusion Collectively, our data do not support the hypothesis that xylitol accumulation occurs primarily due to differences in cofactor preferences between xylose reductase and xylitol dehydrogenase; instead, gene expression plays a major role in response to oxygen levels. We have also identified the yeast Sc. xylosifermentans as a potential source for genes that can be engineered into S. cerevisiae to improve xylose fermentation and biofuel production.

BACKGROUND:Colletotrichum fungi infect a wide diversity of monocot and dicot hosts, causing diseases on almost all economically important plants worldwide. Colletotrichum is also a suitable model for studying gene family evolution on a fine scale to uncover events in the genome associated with biological changes. RESULTS:Here we present the genome sequences of 30 Colletotrichum species covering the diversity within the genus. Evolutionary analyses revealed that the Colletotrichum ancestor diverged in the late Cretaceous in parallel with the diversification of flowering plants. We provide evidence of independent host jumps from dicots to monocots during the evolution of Colletotrichum, coinciding with a progressive shrinking of the plant cell wall degradative arsenal and expansions in lineage-specific gene families. Comparative transcriptomics of 4 species adapted to different hosts revealed similarity in gene content but high diversity in the modulation of their transcription profiles on different plant substrates. Combining genomics and transcriptomics, we identified a set of core genes such as specific transcription factors, putatively involved in plant cell wall degradation. CONCLUSIONS:These results indicate that the ancestral Colletotrichum were associated with dicot plants and certain branches progressively adapted to different monocot hosts, reshaping the gene content and its regulation.

Neural Machine Translation (NMT) is a prominent natural language processing technique that is being used to develop conversational AI technology. Traditional industrial chatbots often rely on scripted responses and lack the ability to provide real-time data-driven interactions. Developing advanced AI chatbots like Google Bard or ChatGPT are out of reach for smaller or medium-sized organizations due to their scale and associated costs. Chatbots are majorly restricted by the data on which they were trained on and have no knowledge of current events. This research project intends to research and develop an approach that enables chatbots to provide live and up-to-date information in their responses and can be developed in minimalistic costs so that even smaller or medium-level organizations can afford to provide interactive AI chatbots. We experiment with various techniques in terms of the type of data being used to harness live capabilities. We focus on optimizing the hyperparameters required for building a conversational AI agent and leverage open-source technologies to minimize costs. To ensure flexibility and affordability, we adopt a microservice architecture that combines Attention based NMT models and Transformer Models with live API services features, leveraging the RASA actions API. This approach allows us to develop a prototype of an advanced chatbot that goes beyond the traditional scripted responses by providing real-time information to users and being affordable to develop.

Generally candidates apply to multiple jobs with a single resume and do not tend to customize their resume to match the job description. This hampers their chances of getting a resume shortlisted for the job. The project aims to help such candidates build job tailored resumes that help them create a customized and targeted resume for a specific job. The tool specifically targets candidates’ employment work history for resume content generation. We create a synthetic dataset built from candidates’ employment history and online job descriptions. We use natural language processing (NLP) techniques to extract and organize the dataset, experiment with multiple dataset variations and cite ways to effectively build the dataset for the proposed task. We then use natural language generation by fine tuning GPT-2 for the task of resume content generation. Finally we evaluate the fine tuned model on various metrics and report our findings.

The internal workings of deep learning models are opaque and are considered as black boxes due to their nested and non-linear structure. This opaque nature of the deep neural networks makes it difficult to interpret the reason behind their output, thus reducing trust and verifiability of the system where these models are applied. This paper explains a systematic approach to identify the clusters with most misclassifications or false label annotations. For this research, we extracted the activation vectors from a deep learning model, DNABERT, and visualized them using t-SNE to decode the reason behind the results that are produced. We applied K-means in a hierarchical fashion on the activation vectors for a set of training instances. We analyzed cluster mean activation vectors to find any patterns in the errors across K-means clusters. The cluster analysis revealed that the predictions were uniform, or nearly 100 percent same, in clusters of similar activation vectors. It was found that two clusters containing most of their objects belonging to the same true class tend to be closer together than clusters of opposite classes. The means of objects of the same true label are closer if two clusters have the same predicted labels rather than opposite predicted labels, showing that the activation vectors reflect both predicted and true classes. We propose a heuristic to find the clusters with a high number of misclassifications or incorrect label annotations using the between clusters and within clusters mean vector analysis. This can aid in identifying misclassifications of DNA sequences or problems with sequence tagging.

The mutualistic ectomycorrhizal (ECM) fungal genus Pisolithus comprises 19 species defined to date which colonize the roots of >50 hosts worldwide suggesting that substantial genomic and functional evolution occurred during speciation. To better understand this intra-genus variation, we undertook a comparative multi-omic study of nine Pisolithus species sampled from North America, South America, Asia, and Australasia. We found that there was a small core set of genes common to all species (13%), and that these genes were more likely to be significantly regulated during symbiosis with a host than accessory or species-specific genes. Thus, the genetic “toolbox” foundational to the symbiotic lifestyle in this genus is small. Transposable elements were located significantly closer to gene classes including effector-like small secreted proteins (SSPs). Poorly conserved SSPs were more likely to be induced by symbiosis, suggesting that they may be a class of protein that tune host specificity. The Pisolithus gene repertoire is characterized by divergent CAZyme profiles when compared with other fungi, both symbiotic and saprotrophic. This was driven by differences in enzymes associated with symbiotic sugar processing, although metabolomic analysis suggest that neither copy number nor expression of these genes is sufficient to predict sugar capture from a host plant or its metabolism in fungal hyphae. Our results demonstrate that intra-genus genomic and functional diversity within ECM fungi is greater than previously thought, underlining the importance of continued comparative studies within the fungal tree of life to refine our focus on pathways and evolutionary processes foundational to this symbiotic lifestyle.