Biology subjects

Walker, A.

Publications and source records attributed to Walker, A..

At least 19 recordsLinked to original sources

bioRxiv2026

NM-RARe: A FRAMEWORK FOR REGIONAL ARTHROPOD CONSERVATION ASSESSMENT: INSIGHTS FROM NEW MEXICO

Arthropods represent the most species-rich animal group on the planet, yet they remain largely overlooked in formal conservation planning. New Mexico is a US state with exceptional biodiversity, including numerous endemic, rare, and threatened arthropod species. Recent attention to global insect declines has highlighted the need to document possible threats to and declines of New Mexico insects and other arthropods. All terrestrial arthropods in New Mexico lacked state-level regulatory protection until 2025, and baseline data on the taxonomy, ecology, geographic distribution and conservation needs for most New Mexico arthropod species remained absent. To address this gap, we developed the New Mexico Rare Arthropod Resource (NM-RARe), a publicly accessible regional database for arthropod species and subspecies of conservation concern in New Mexico. Drawing from multiple vetted sources, we compiled an initial working list of 7,018 taxa and conducted 261 preliminary IUCN Red List assessments and 185 NatureServe state rank assessments, covering 285 total taxa. Of those assessed by IUCN Red List standards, 52 were subspecies and 209 were species level assessments, 165 of which met the criteria for inclusion and were incorporated into the NM-RARe database (https://nmrare.org/). Ninety-three of the pollinating insect species from NM-RARe were then formally incorporated into New Mexico's 2025 State Wildlife Action Plan (SWAP), representing a significant conservation milestone. Beyond its immediate regional impact, NM-RARe presents a standardized, scalable methodology adaptable to any region or taxonomic group, offering a replicable framework for arthropod conservation assessment worldwide.

ecology↗

bioRxiv2026

A foundation model enables prediction of natural product molecular properties, bioactivity, and structural similarity from biosynthetic gene cluster sequence

Genome mining is a powerful technique in natural product discovery, where biosynthetic gene clusters that are likely to produce novel or desirable natural products are identified through bioinformatic analysis. There are many more predicted biosynthetic gene clusters than can easily be experimentally characterized. Additional computational methods to prioritize biosynthetic gene clusters by the bioactivity, structural properties, or novelty of the product would make genome mining more efficient. Multiple machine learning/artificial intelligence models have been developed to predict product properties from biosynthetic gene cluster sequence, but they are limited by small quantities of training data. Model pretraining with unlabeled data is a powerful technique to develop models that can learn on a limited amount of labeled training data. Biosynthetic gene clusters are well suited to this strategy because there are many predicted clusters with only a small percentage being characterized. This paper reports BGC-MLM, a foundation model that is pretrained with a masked language task on predicted biosynthetic gene clusters and then fine-tuned for downstream applications including prediction of product structural class, bioactivity, chemical properties, counts of functional groups, and chemical fingerprint. Comparison to a model trained without pretraining shows that pretraining generally improves performance. BGC-MLM shows better or similar performance to existing specialized methods for these tasks, demonstrating its utility as a foundation model for natural product genome mining.

bioinformatics↗

bioRxiv2026

GYDE: A collaborative drug discovery platform for AI-powered protein design and engineering

As computational tools and machine learning models for protein sciences continue to advance and proliferate, bench scientists face increasing technical challenges adopting these tools for specific applications such as drug discovery. Here we present GYDE (Guide Your Design and Engineering), an open-source, versatile, and web-based collaboration platform designed to make computational analyses of proteins and antibodies easily accessible to bench scientists. GYDE enables the exploration of sequence-structure-function relationships through a tightly integrated visual interface, offering researchers a comprehensive exploration of protein functional determinants either via real assay data or computational tools. GYDEs intuitive interface facilitates seamless access to cutting-edge AI models for protein and antibody structure prediction, design, and downstream analyses. The flexible and easy addition of new tools and models is facilitated by the use of the Slivka compute API. The platform supports saved sessions that enable researchers to easily share their findings with other users, fostering a more collaborative scientific community. GYDE is freely available for protein scientists in academia and industry to build drug discovery analytics platforms customized to their needs.

bioinformatics↗

bioRxiv2026

Multimodal analysis of cell-free DNA identifies epigenetic biomarkers for amyotrophic lateral sclerosis diagnosis and progression

The role of the epigenome in age-related neurodegenerative disorders remains understudied. Here, we analyzed circulating cell-free DNA (cfDNA) from blood to detect methylation changes as a liquid-biopsy for Amyotrophic Lateral Sclerosis (ALS). Our study included 20 patients with sporadic ALS, 10 patients with C9orf72-associated ALS, 10 asymptomatic carriers of the C9orf72 repeat expansion mutation, and 21 non-disease controls. Following targeted enzymatic methyl-sequencing (EM-seq) of [~]4 million CpG sites, we detected numerous differentially methylated genes, including several implicated in ALS disease risk and pathogenesis. By integrating multiple epigenetic features, we delineated a distinct epigenetic signature, which achieved an average area under the curve (AUC) of 0.91 {+/-} 0.10 upon receiver operator characteristic (ROC) analysis, which enabled detection of [~]70% of ALS patients with close to 100% specificity. Furthermore, we also identified a set of genes whose methylation status significantly correlated with clinical disease progression and cerebrospinal fluid (CSF) neurofilament levels. Our results reveal the potential of cfDNA-based biomarkers to accurately diagnose ALS and potentially predict disease progression.

genetics↗

bioRxiv2026

Identification of Siderophores with Unexpected Antibacterial Properties from Actinoplanes teichomyceticus

Actinoplanes teichomyceticus is a well-established producer of bioactive secondary metabolites, including the glycopeptide antibiotic teicoplanin. Although its antibiotic biosynthetic capacity has been extensively investigated, its siderophore diversity and any additional biological functions of these iron-chelating metabolites remain comparatively underexplored. We identified a reproducibly bioactive, teicoplanin-independent fraction that inhibited Bacillus spizizenii. Molecular networking applied to this fraction identified hydroxamate ferrioxamine and desferrioxamine-type siderophores as the dominant metabolites, including acylated analogs detected as Al3+- and Fe3+-chelated species. Robust siderophore secretion was confirmed by the CAS assay. Notably, siderophore-enriched fractions exhibited selective antibacterial activity against Gram-positive bacteria, with minimum inhibitory concentrations of approximately 16 {micro}g/mL against B. spizizenii and partial inhibition of Staphylococcus aureus, while no activity was observed against Escherichia coli. Synthetic C7 and C9 acyl-desferrioxamine analogs showed enhanced antibacterial activity upon Al3 chelation, indicating a metal-dependent bioactivity. These findings reveal an unexpected antibacterial role for ferrioxamine-type siderophores produced by A. teichomyceticus, extending their function beyond iron acquisition, possibly through a "Trojan horse" (or "Trojan metal") mechanism.

biochemistry↗

bioRxiv2026

Microbial metabolism of methotrexate produces a STAT3 signaling molecule that alleviates gut inflammation

Methotrexate (MTX) therapy in inflammatory bowel disease (IBD) is often limited by inter-individual variability in clinical response and adverse effects. Gut microbiota contribute to MTX therapeutic response and toxicity by metabolizing MTX and altering its bioavailability. However, how inflammation alters microbial MTX metabolism and how its metabolites influence the host remain poorly understood. Here, we identify Clostridium asparagiforme as a potent and efficient metabolizer of methotrexate, producing deoxyaminopteroic acid (DAMPA) in the distal gastrointestinal tract. We demonstrate that DAMPA preserves mitochondrial integrity by promoting mitophagy in intestinal epithelial cells through mitochondrial STAT3 signaling. DAMPA administration attenuates intestinal inflammation in vivo, and improves metabolic dysfunction associated with IBD. Together, these findings reveal an unappreciated role for a gut microbial MTX metabolite in mediating epithelial homeostasis during intestinal inflammation, thus reframing microbial MTX metabolism from passive drug detoxification to active regulation of host mitochondrial and inflammatory homeostasis.

microbiology↗

bioRxiv2026

PIP2;1 aquaporin promotes early stomatal closure in grapevine leaves during water stress

Aquaporins play a key role in plant responses to drought. Our previous work showed limited embolism in grapevine leaves under mild water stress and suggested that the outside-xylem water pathway plays a dominant role in reducing leaf hydraulic conductance (Kleaf) during dehydration. We used CRISPR-Cas9 to knockout the PIP2;1 aquaporin encoding gene in Vitis vinifera cv. Chardonnay to study how leaf function during dehydration is affected by this aquaporin isoform. We measured functional responses like stomatal and photosynthetic responses as well as Kleaf to compare wild-type and two independent PIP2;1 knockout lines. Under moderate drought, mutants maintained greater stomatal conductance (gs) and photosynthetic rates as {Psi}w declined. No significant differences were observed in mesophyll conductance (gm) across genotypes, however, mutants exhibited slightly higher values under moderate drought. Interestingly, all lines exhibited similar Kleaf vulnerabilities to drought. Our findings show that PIP2;1 induces earlier stomatal closure during dehydration while not modulating Kleaf responses across genotypes. This rapid response in WT plants would prevent further water loss that would lead to higher xylem tensions that can lead to embolism. These findings show that multiple mechanisms collectively limit leaf gas exchange and water loss during dehydration, enhancing our understanding of plant resilience to changing environments.

plant biology↗

bioRxiv2025

Long-Term Intestinal Epithelial Remodeling Induced by Acute Protein-Energy Malnutrition

Protein-energy malnutrition (PEM) is a global health burden with lasting effects that extend well beyond the initial nutrient deficiency. To systematically investigate the long-term effects of a single episode of PEM on the structure and function of the intestinal epithelium and its associated microbiota, we employed a comprehensive multi-omics approach, including (spatial) transcriptomics, DNA methylation analysis, fecal metagenomics, and metabolomics. Our findings show that PEM persistently alters the intestinal epithelium by depleting Paneth cells and suppressing antimicrobial gene expression - changes linked to DNA methylation that persist despite dietary recovery. In germ-free mice, the sustained epithelial phenotype after was absent. We identified the microbial lipid metabolite 9-HODE and epigenetically deregulated PPAR-driven GDF15 expression as key molecular drivers of the persistent PEM-induced Paneth cell dysfunction. Targeting microbial lipid production and its link to the host GDF15 pathway could offer novel therapeutic strategies for long-term consequences of malnutrition and other Paneth cell-associated diseases.

immunology↗

bioRxiv2025

Diagnosing linearity along the carbon cascade in terrestrial biosphere models

Elevated carbon dioxide (eCO2) fertilises photosynthesis, driving an increase in terrestrial gross primary production (GPP). However, it is unclear how effectively increased GPP propagates along the "carbon (C) cascade" to increase net primary production (NPP) and vegetation C stocks (Cveg) in different plant compartments. Vegetation models simulate divergent C cycle projections and have been criticised for being overly photosynthesis-driven (source-driven), neglecting processes that lead to non-linear behaviour in response to the GPP increase, which may attenuate (or amplify) changes in NPP and vegetation C stocks. Here, we introduce an analytical framework to diagnose linearity (L) of the land C cycle as the ratio of relative changes in linked fluxes and pools and apply it to outputs from 16 models of the TRENDY v11 ensemble. We found widely varying patterns in L across models and for the different links. Six models showed a clear dominance of larger relative changes in NPP than in GPP in global simulations (LNPP:GPP >1 for >60% of gridcells), indicating increased carbon use efficiency under eCO2. Only three models had LNPP:GPP < 1 for >60% of gridcells. Four models showed a clear dominance of larger relative changes in steady-state Cveg than in NPP, while five models showed an opposite pattern - in both cases with a large spread of LCveg*:NPP across gridcells within models. Three models showed a larger relative increase in root C than in Cveg, while two models showed a clear dominance of the opposite pattern. Widely differing distributions of L across models and links reveal a strong influence of alternative process representations (nonlinear behaviour) in individual models. However, for all links, L deviations from 1 were roughly balanced across the model ensemble, leading to an overall linear behaviour of terrestrial C cycle representations.

ecology↗

bioRxiv2025

Multi-omics analysis highlights the link of aging-related cognitive decline with systemic inflammation and alterations of tissue-maintenance

Aging-related cognitive decline is associated with changes across different tissues and the gut microbiome, including dysfunction of the gut-brain axis. However, only few studies have linked multi-organ alterations to cognitive decline during aging. Here we report a multi-omics analysis integrating metabolomics, transcriptomics, DNA methylation, and metagenomics data from hippocampus, liver, colon, and fecal samples of mice, correlated with cognitive performance in the Barnes Maze spatial learning task across different age groups. We identified 734 molecular features associated with cognitive rank within individual data layers, of which 227 features remain when integrating all data layers with each other. Among the single-layer predictors, several host and microbial features were highlighted, with host-associated markers being predominant. Host features associated with cognitive function mainly belong to innate and adaptive inflammatory activity (inflammaging) and developmental processes. Our findings suggest that cognitive decline in aging is tightly coupled to systemic, age-associated inflammation, potentially initiated by microbiome-driven gastrointestinal inflammatory activity, emphasizing a link between peripheral tissue alterations and brain function.

systems biology↗

bioRxiv2025

Sodium Butyrate Rescues Neurodevelopmental Deficits Following Perinatal Methadone Exposure

Prenatal opioid exposure (POE) induces long-term neurodevelopmental, behavioral and cognitive deficits for which no targeted treatments are available. The mechanisms underlying POE deficits are poorly understood, but have been linked to a range of central, peripheral, and enteric nervous system changes. Emerging evidence indicates that maternal microbiota changes may also contribute to these long-term deficits in offspring. Here we test the efficacy of the short-chain fatty acid sodium butyrate (NaB) to mitigate POE-induced deficits in a rat model. Both methadone and sodium butyrate treatments altered dam microbiota composition and function: notably methadone disrupted dam gene expression of microbial enzymes critical for butyric acid production and reduced faecal butyric acid levels. In postnatal day 9 pups, methadone increased gut barrier permeability that was reversed with NaB, and enzymatic disruptions were observed in pups at postnatal day 21 that resolved in adulthood. POE induced anxiety-like behavior in adolescence, and adult deficits in working spatial memory and attentional processing that were partially rescued in rats that had received prenatal NaB. POE was associated with decreased myelination in the hippocampus, and this was partially reversed by NaB. Together these results highlight for the first time the link between the gut-brain axis in animal models of POE. Furthermore, they provide the first indication in a rat model of NaB as a simple yet effective treatment to significantly improve the outcomes of children born with POE.

animal behavior and cognition↗

bioRxiv2025

Tissue scarring provides a biomechanical framework to promote mammalian bile duct regeneration through the activation of integrin-SRC/FAK signalling.

Following chronic injury, the adult mammalian bile duct regenerates by forming new branches, essentially replumbing the ductular system to overcome blockages and breaks. To regenerate effectively, biliary epithelial cells (BECs) receive a range of pro-mitogenic signals from myofibroblasts, which concurrently deposit a collagen-rich scar around the duct as it regrows. Despite epithelial regeneration and scarring occurring side-by-side, whether the deposition of scar tissue regulates ductular regeneration per se remains unclear. By inducing ductular fibrosis and regeneration in vivo, we show that the formation of collagen-I-rich scars around regenerating ducts changes the local biomechanical properties of these tissues, promoting the growth of ducts. Critically, this changing structural landscape is perceived by a spatially restricted population of biliary epithelial cells which forms a "leading-tip" of integrin-2-high cells. This leading-tip undergoes partial-EMT-type reprogramming, allowing it to become migratory and coordinate ductular regeneration. We show that this process is directly driven through an integrin-2-SRC/FAK signalling axis; thereby connecting epithelial regeneration directly to the changing fibrotic environment in chronic ductular disease. HighlightsO_LIChronic liver disease results in the formation of stiff, collagen scars around ducts. C_LIO_LINew ducts acquire high levels of integrin-2 which is spatially localised to a "leading-tip", which loses epithelial features. C_LIO_LIIntegrin-2{beta}1-SRC/FAK signalling regulates ductular migration by linking changes in the bio-structural composition of the liver with ductular cells. C_LI

cell biology↗

bioRxiv2025

AI-assisted isolation of bioactive Dipyrimicins from Amycolatopsis azurea and identification of its corresponding dip biosynthetic gene cluster

One of the major challenges in natural product discovery is the prioritization of compounds with useful activities from microbial sources. In particular, this is a challenge in genome mining for novel natural products, where the structures and activities of compounds produced by bioinformatically identified and uncharacterized biosynthetic gene clusters remain unknown. Here, we utilize a machine learning model to predict the antibacterial activity of a natural product from its biosynthetic gene cluster (BGC). We prioritized the strain Amycolatopsis azurea DSM 43854 which was predicted by machine learning to have the capacity to produce multiple natural products with antibacterial activity. Together with bioactivity-guided fractionation, we isolated dipyrimicins A and B from Amycolatopsis azurea DSM 43854 and, for the first time, linked them to their BGC. This dip BGC was predicted by our model to encode a product with 75% antibacterial probability and shares only 40-52% similarity with previously characterized BGCs. We confirmed the antimicrobial properties of the dipyrimicins against a few test strains and identified key tailoring enzymes, including an O-methyltransferase and amidotransferase, that differentiated them from other related 2,2-bipyridine biosynthetic pathways. Importantly, As the dip BGC was not in the training set of the model, our results demonstrate the ability of the model to generalize beyond its training set and the potential of machine learning to accelerate novel bioactive natural product discovery and deorphanization of biosynthetic gene clusters.

biochemistry↗

bioRxiv2025

The carnosinase dipeptidase CNDP1 is a novel metabolic vulnerability in brain metastasis

Brain metastatic cells undergo metabolic adaptations, such as increased reliance on oxidative phosphorylation. Integrating proteomic and transcriptomic profiling of patient samples revealed a consistent upregulation of the carnosine dipeptidase-1 (CNDP1) in brain versus extracranial metastases. Carnosine is an abundant metabolite in brain and muscle, known to exert anti-proliferative effects on cancer cells. Here, we demonstrate that CNDP1 inhibition suppresses both the establishment and maintenance of melanoma brain metastasis while its ectopic expression is sufficient to confer brain metastatic potential to poorly metastatic cells. CNDP1 suppression results in activation of the Integrated Stress Response via Heme-Regulated Inhibitor Kinase and reprogrammed translation towards preferential expression of mitochondrial and survival transcripts. We further show that CNDP1 upregulation supports mitochondrial activity by limiting the levels of its substrate carnosine, a copper ionophore, thus protecting metastatic cells from carnosine-induced copper toxicity. Our studies reveal a novel metabolic adaptation during brain metastasis, which can be leveraged for therapeutic purposes.

cancer biology↗

bioRxiv2025

Synaptic changes contribute to persistent extra-motor behaviour deficits in the rNLS8 TDP-43 mouse model of amyotrophic lateral sclerosis

Extra-motor symptoms are increasingly recognised in amyotrophic lateral sclerosis (ALS), encompassing cognitive, social, and behavioural deficits that can substantially impact quality of life. TAR DNA binding protein 43 (TDP-43) pathology is the central disease marker of almost all cases of ALS and approximately half of all frontotemporal dementia (FTD). However, the mechanisms linking TDP-43 pathology with extra-motor symptoms in TDP-43-associated neurodegenerative diseases remain unresolved. In this study, we used the rNLS8 mouse model, which expresses human TDP-43 with an ablated nuclear localisation sequence (hTDP-43∆NLS) in a doxycycline-regulatable manner causing progressive motor decline reminiscent of ALS, to delineate the molecular changes associated with disease-relevant phenotypes. We found that in addition to previously reported dramatic motor decline, rNLS8 mice also develop extra-motor phenotypes consistent with FTD, including disinhibition-like and anxiety-like behaviours, and social interaction impairments. These changes began in the earliest disease stages and remained readily detectable even when rNLS8 mice became severely motor impaired. Notably, extra-motor deficits persisted in rNLS8 mice that had recovered motor function upon hTDP-43∆NLS transgene suppression, regardless of whether recovery was initiated at timepoints prior to or after overt neurodegeneration begins. Transcriptomic analysis of rNLS8 mouse cortex tissues revealed early alterations in expression of 321 genes, most notably involving neuroinflammatory-related pathway activation, and all but 2 of these genes returned to control levels upon suppression of hTDP-43∆NLS expression. Further, 814 genes showed differential exon usage, indicating changes in alternative splicing, in rNLS8 mouse cortex. Of these, differential exon usage of 10 neuronal genes persisted after hTDP-43∆NLS transgene suppression, including synapse component genes Nrxn1, Unc13a, and Gls. Similarly, proteomics analysis of the cortex of rNLS8 mice revealed depletion of synaptic proteins, particularly those involved in glutamatergic signalling pathways, which also persisted following hTDP-43∆NLS transgene suppression. Similar glutamatergic pathway changes were detected in human ALS and FTD post-mortem cortex tissues. Our findings indicate that extra-motor phenotypes emerge early in disease in rNLS8 mice and remain evident despite progressive motor impairments. Further, extra-motor phenotypes persist even upon motor recovery, correlating with specific synaptic gene expression and splicing changes. Overall, this study suggests the potential utility of an expanded suite of behavioural paradigms in preclinical testing using rNLS8 mice, with enhanced relevance to the diversity of TDP-43 proteinopathies including FTD. Our findings further suggest that targeting glutamatergic synaptic components may be an avenue to correct extra-motor deficits associated with TDP-43 pathology.

neuroscience↗

bioRxiv2025

Targeting deubiquitinating enzymes (DUBs) and ubiquitin pathway modulators to enhance host defense against bacterial infections

The rise of antibiotic-resistant bacterial pathogens poses a critical global health challenge, necessitating innovative therapeutic approaches. This study explores host-targeted therapies (HTTs) by focusing on deubiquitinating enzymes (DUBs), essential modulators of the ubiquitin-proteasome system (UPS) that regulate host-pathogen interactions during many bacterial infections. Using Salmonella-infected macrophages as a model, we identified UPS modulators that enhance bacterial clearance and observed significant changes in DUB expression, particularly USP25, USP46, and Otud7b. The small-molecule DUB inhibitor AZ-1 significantly reduced intracellular bacterial loads in vitro and mitigated early disease severity in a murine model by decreasing fecal bacterial loads and preserving host weight. However, AZ-1 alone did not achieve complete clearance of Salmonella and required combination with extracellular-targeting antibiotics for optimal efficacy. Notably, AZ-1 demonstrated broad-spectrum activity against multidrug-resistant pathogens, including Pseudomonas aeruginosa, Klebsiella pneumoniae, and Acinetobacter baumannii. Transcriptomic analyses revealed infection-induced DUB regulation and highlighted pathways modulating immune responses, including TNF- secretion. These findings highlight the potential of targeting the UPS as a host-directed antimicrobial strategy and provide a foundation for developing innovative therapies to combat antimicrobial resistance.

microbiology↗

bioRxiv2024

PMS2 has both pro-mutagenic and anti-mutagenic effects on repeat instability in the Repeat Expansion Diseases

Expansion of a disease-specific tandem repeat is responsible for >45 Repeat Expansion Diseases (REDs). The expansion mutation in each of these diseases has different pathological consequences and most are currently incurable. If the underlying mechanism of mutation is shared, a strategy that slows repeat expansion in one RED may be applicable to multiple REDs. However, the fact that PMS2, a component of the MutL mismatch repair complex, promotes expansion in some models and protects against it in others suggests that the expansion mechanisms may differ. We show here using mouse models of two REDs caused by different repeats that PMS2 has similar effects in both models, with the loss of PMS2 resulting in an increase in expansions in some tissues and a loss of expansion in others. This is consistent with a protective effect of PMS2 in the first case and a role in promoting expansion in the second. Furthermore, we show in mouse embryonic stem cells that lower levels of PMS2 promote expansion while higher levels protect against it, with the ability to promote expansion depending on the PMS2 nuclease domain. Our findings lend support to the hypothesis that REDs share a common expansion mechanism and provide insights into the processes involved. Significance statementCollectively the Repeat Expansion Diseases (REDs) represent a significant health burden. Since the consequences of the expansion mutation differ across diseases, therapeutic approaches that block the underlying mutation are appealing, particularly if the mechanism is shared. However, the conflicting effects of PMS2 loss in different RED models challenges this idea. Here we show using two different RED models that these disparate effects can be reconciled into a single model of repeat expansion, thus increasing confidence that the REDs do all share a common mutational mechanism.

genetics↗

bioRxiv2024

Assessing the ability of ChatGPT to extract natural product bioactivity and biosynthesis data from publications

Natural products are an excellent source of therapeutics and are often discovered through the process of genome mining, where genomes are analyzed by bioinformatic tools to determine if they have the biosynthetic capacity to produce novel or active compounds. Recently, several tools have been reported for predicting natural product bioactivities from the sequence of the biosynthetic gene clusters that produce them. These tools have the potential to accelerate the rate of natural product drug discovery by enabling the prioritization of novel biosynthetic gene clusters that are more likely to produce compounds with therapeutically relevant bioactivities. However, these tools are severely limited by a lack of training data, specifically data pairing biosynthetic gene clusters with activity labels for their products. There are many reports of natural product biosynthetic gene clusters and bioactivities in the literature that are not included in existing databases. Manual curation of these data is time consuming and inefficient. Recent developments in large language models and the chatbot interfaces built on top of them have enabled automatic data extraction from text, including scientific publications. We investigated how accurate ChatGPT is at extracting the necessary data for training models that predict natural product activity from biosynthetic gene clusters. We found that ChatGPT did well at determining if a paper described discovery of a natural product and extracting information about the products bioactivity. ChatGPT did not perform as well at extracting accession numbers for the biosynthetic gene cluster or producers genome although using an altered prompt improved accuracy.

bioinformatics↗