Ancient DNA (aDNA) sequences present unique challenges for taxonomic classification due to extreme fragmentation (reads 20-100 bp), end-biased cytosine deamination, and high contamination rates. Conventional metagenomic classifiers based on exact k-mer matching or alignment lose discriminative power on such short and damaged reads, limiting the analysis of paleogenomic samples. We present FALCON2, a compression-based metagenomic classifier that leverages position-aware finite-context models to maintain high accuracy on degraded viral ancient viruses. FALCON2 consolidates the capabilities of its predecessor, FALCON-meta, into a unified executable with enhanced features including model persistence, direct processing of compressed inputs, multiple file handling, and optional pre-filtering methodologies for contaminated samples. Under controlled benchmarking with database, taxonomy, and thread parity on simulated viral datasets, FALCON2 achieved an Area Under the Curve of Receiver Operating Characteristic (AUC-ROC) of 0.999, an Area Under Precision-Recall Curve (AUPRC) of 0.968, and an F1-score of 0.918, substantially outperforming Centrifuge (AUPRC = 0.625), Kraken2 (AUPRC = 0.184), and CLARK-S (AUPRC = 0.013) on pooled micro-averaged metrics. FALCON2's advantage is most pronounced on ultra-short reads (20-40 bp), where exact k-mers become sparse. FALCON2 pre-filtering at threshold 0.7 improved precision by 10 percentage points with negligible recall loss. FALCON2 runs on systems with 4-8 GB RAM for typical analyses. FALCON2 is freely available at https://github.com/cobilab/FALCON2 under GPL v3 license. Benchmarking data and scripts are archived at DOI: https://doi.org/10.5281/zenodo.17291214. Supplementary data are available at Bioinformatics online.
Aspergillosis is a respiratory disease in birds, particularly falcons, often diagnosed late due to nonspecific signs and limited performance of current diagnostic tools. This study aimed to identify novel plasma biomarkers for aspergillosis comparing plasma samples from 15 healthy falcons and 15 affected falcons (9 with early and 6 with advanced aspergillosis). First, the presence of the causative agent, Aspergillus fumigatus, was ruled out through a proteotyping strategy. Proteomic profiling using data-independent acquisition mass spectrometry further identified 861 avian proteins. In advanced infection, 50 host proteins were significantly modulated in terms of abundance, including acute-phase proteins such as the haptoglobin isoform X2, alpha-1-acid glycoprotein, and serum amyloid A-like protein. Several immunoglobulin isoforms and functionally uncharacterized proteins were also differentially abundant. In contrast, no significant changes were observed between healthy and early affected birds. Comparison between early and advanced cases revealed nine proteins with significant abundance shifts, suggesting that they could be interesting disease progression markers. These findings highlight specific host protein responses in advanced aspergillosis animals and support the potential of targeted monitoring of plasma biomarkers for earlier, non-invasive diagnosis in falcons. Further validation on a larger cohort is warranted to assess clinical relevance and diagnostic performance in veterinary medicine. SIGNIFICANCE: Aspergillosis remains one of the most significant infectious diseases affecting birds, where delayed diagnosis frequently results in fatal outcomes. Current noninvasive diagnostic is still challenging due to non-specific clinical signs and limited sensitivity and specificity of current diagnostic tests. This study presents a comprehensive proteomic characterization of falcon plasma during aspergillosis, taking advantage of recent advances in high-resolution data-independent acquisition mass spectrometry and the availability of annotated falcon genomes. Comparison of healthy, early, and advanced clinical stages revealed 50 significantly dysregulated proteins, including key acute-phase proteins such as haptoglobin, serum amyloid A, alpha-1-acid glycoprotein, ceruloplasmin, and immunoglobulin light chains. These alterations reflect systemic inflammatory and immune responses to infection and highlight promising biomarkers for early, noninvasive diagnosis. This work contributes to the development of protein-based diagnostic assays, thereby improving clinical management and welfare of captive and wild birds, while also expanding proteomic resources for non-model avian species.
Presence of intravenous contrast on computed tomography (CT) scans is often unreliably documented, especially in large research datasets. FALCON is an open-access fully automated deep learning model enabling large-scale intravenous contrast detection and body part classification for CT scans of the head and neck (HN), chest, abdomen, and pelvis (AP). This study used six independent datasets consisting of 3138 CT scans of the HN, chest, and AP of 3126 patients from five institutions between 1996 and 2023 to train and validate four CNN models for intravenous contrast detection and body part classification. The ground truth of intravenous contrast presence was verified by a radiologist. We used ResNet9 network architecture and integrated the four models into a graphical user interface. We assessed FALCON's performance with F1 scores and compared FALCON's annotation time to manual annotation by human experts. In the external test set containing 1348 scans, the F1 score for intravenous contrast detection was 99.4% (95%CI: 98.8, 99.9) for HN CT, 98.3% (95%CI: 96.9, 99.5) for chest CT, and 98.1% (95%CI: 96.9, 99.1) for AP CT. The F1 score for body part classification alone on unseen data was 100% for HN, chest, and AP CT. Compared to human experts, annotation of a single scan with FALCON required 1.3 s vs. 21 s for HN CT, 1.8 s vs. 33 s for chest CT, and 3.7 s vs. 1.6 s for AP CT. The open-access FALCON model ( https://github.com/FintelmannLabDevelopmentTeam/Falcon ) quickly and reliably detects intravenous contrast and classifies body part on CT scans.
Climate change is considered a key driver for shaping ecological and evolutionary processes of Arctic animals. Historical glaciation has profoundly influenced the distribution and genetic differentiation of Arctic vertebrates, and recently Arctic species are facing new and intensifying threats from rapid global warming. Understanding how past, recent and future climate change has, and will influence the evolution of Arctic animals is, therefore, crucial for effective conservation planning. Here we combine whole-genome sequencing, demographic inference, and species distribution modeling (SDM) to assess the eco-evolutionary responses of the gyrfalcon (Falco rusticolus), a resident Arctic apex predator, to climate change. Assembling a genome reference and using samples from three breeding regions across the Eurasian Arctic (Kola, Yamal, and Chukotka peninsulas), we found genetic differentiation of gyrfalcon populations from west to east, that arose during the late Pleistocene (12.9-14.7 thousand years ago (ka)) and subsequently persisted in isolation, until gene flow into the Yamal population resumed during the late Holocene. The extant gyrfalcon populations exhibit low genetic diversity, elevated inbreeding coefficients, and high genetic loads compared to the closely related saker falcon (Falco cherrug), and some other threatened species with small populations, likely due to a population bottleneck about 1 ka, which might compromise the long-term viability of this Arctic raptor. Additionally, the effective population size (Ne) of the Kola gyrfalcon population was inferred to be in decline over the past 165-60 years. SDM based on ensemble models further predicts a substantial reduction of climatically suitable areas for Kola gyrfalcons under future global warming scenarios. Our study highlights how past climatic fluctuations and ongoing warming jointly shape the genomic landscape of endemic Arctic birds and provides insights into making conservation strategies for Arctic animals in a rapidly warming environment.
The Peregrine Falcon (Falco peregrinus) is a species of conservation interest throughout much of eastern North America, and management efforts for the species are widespread. Peregrines are at risk for exposure to Trichomonas spp. because of their tendency to take pigeons and doves (Columbiformes) as prey. We investigated the prevalence of Trichomonas spp. in wild nestling Peregrine Falcons banded in nests in Kentucky, USA. We tested throat swabs collected from 266 Peregrine Falcons during 2005-23. A total of 21 birds tested positive for Trichomonas spp.; none of the falcons that tested positive were resighted as adults. We treated a subset of nestlings for Trichomonas spp. infection, including two in the nest (without removal), two nestlings temporarily removed from the nest for treatment, and 10 placed with wildlife rehabilitators for treatment. Rehabilitated nestlings were released at an age of ≥75 d. However, we did not confirm any benefits to survival from treatment. Trichomonosis may hinder Peregrine Falcon nestling survival in certain areas, especially urban locations.
Background/Objectives: The Saker Falcon (Falco cherrug) is an endangered raptor species of ecological and conservation relevance. Despite its status, data regarding its microbiota and the prevalence of antimicrobial resistance (AMR) remain scarce, especially in Eastern Europe. This single-facility study aims to investigate the phenotypic and genotypic AMR profiles of Gram-negative bacteria isolated from captive Saker Falcons in Western Romania. Methods: Freshly voided fecal droppings were collected non-invasively from 40 clinically healthy Saker Falcons. Bacterial identification was performed using selective media and the VITEK® 2 system. Antimicrobial susceptibility testing (AST) was conducted on a representative subset of 12 isolates. Selected resistance-associated genes were screened by conventional PCR. Results: Escherichia coli was the most prevalent 60% (n = 24/40), followed by Hafnia alvei 10% (n = 4/40) and Pseudomonas spp. 10% (n = 4/40). AST revealed phenotypic resistance among Enterobacteriaceae primarily to ampicillin 20% (n = 2/10), tetracycline 20% (n = 2/10), fluoroquinolones and sulfonamides 10% (n = 1/10), while susceptibility to imipenem 90% (n = 9/10) and gentamicin 90% (n = 9/10) remained high. The targeted resistance-associated genes were detected in selected phenotypically resistant isolates. PCR screening detected blaZ and ampC in 62.5% (n = 5/8) of tested isolates, blaOXA-61 in 37.5% (n = 3/8), blaOXA-51 in 25% (n = 2/8), tetK in 37.5% (n = 3/8), and gyrA in 12.5% (n = 1/8). The isolate used as the negative control, pansusceptible in AST, was confirmed negative for all targeted genes. Conclusions: This single-facility study provides baseline data on AMR traits in Gram-negative bacteria associated with Saker Falcons in Western Romania. Given the limited scale and isolate-based design of the study, the findings should be interpreted cautiously, but they support further investigation of wildlife-associated AMR within a One Health context.
This study focused on the exposure of a terrestrial raptor, the peregrine falcon (Falco peregrinus), in the United Kingdom. In contrast to inland areas, peregrine falcons in coastal areas of North Cornwall, South-west England, have recently declined despite a decreasing trend in environmental legacy organic contaminants. Exposure to per- and polyfluoroalkyl substances (PFAS) is suspected to contribute to one of the causes of this decline. However, unlike studies on aquatic birds, research on PFAS exposure of terrestrial predatory birds remains limited, particularly in British wild birds. To fill this knowledge gap, we have measured PFAS burdens in peregrine eggs from different English areas and compared them with stable isotope and eggshell index values. Our results showed that long-chain perfluoroalkyl acids were predominantly detected in peregrine eggs. Perfluorooctane sulfonyl acid (PFOS), perfluorohexanesulfonic acid (PFHxS), and perfluorooctanoic acid (PFOA) residues significantly differed among counties: PFOS and PFHxS residues were significantly higher in eggs from Devon, the study area around urban settlements, than in Cornwall. PFOA residues were significantly higher in Lancashire, an inland study area, than in Devon. Several perfluoroalkyl carboxylic acid residues showed significant and negative correlations with δ13C in eggs, suggesting that the sources of these PFAS might come from terrestrial habitats. No significant relationship was observed between eggshell index and PFAS residues. Given the variation in PFAS exposure among areas, it remains challenging to determine the impact of PFAS on the Cornwall peregrine population. Further studies are needed to fill these knowledge gaps. The online version contains supplementary material available at 10.1007/s10646-026-03076-x.
Plasmodium huffi was first described in toucans of Brazil and was mistakenly believed to be a species exclusive to the order Piciformes. However, subsequent molecular and morphological studies conducted in red-legged seriemas, along with the diverse range of lineages identified in Anseriformes, Charadriiformes, Cathartiformes, Columbiformes, Galliformes, Pelecaniformes, Struthioniformes, Psittaciformes, and Passeriformes, have revealed that this parasite is, in fact, a generalist. Here, we provide new molecular and morphological data on P. huffi infecting previously unreported avian hosts in Brazil: one black vulture (Coragyps atratus, Cathartiformes, sample size = 29) and one yellow-headed caracara (Daptrius chimachima, Falconiformes, sample size = 6). Additionally, infections were detected in two red-legged seriemas (Cariama cristata, Cariamiformes, sample size = 14), reinforcing the high susceptibility of the species to this parasite. Reports of haemosporidian infections in Cathartiformes are rare. To date, only Haemoproteus catharti, Plasmodium elongatum, and Leucocytozoon toddi have been documented in New World vultures-birds that are widely distributed and abundant across multiple ecosystems. This study marks the first report of haemosporidian infection in black vultures in South America. Although D. chimachima harbors a diverse and abundant haemosporidian community, this is the first record of P. huffi infecting a species within Falconiformes. These findings confirm that P. huffi is highly adapted to a broad range of avian hosts, which can sustain the infection and potentially transmit it to dipteran vectors, as evidenced by the presence of gametocytes in peripheral blood. Additionally, statistical comparisons of the parasite's morphometry across different hosts revealed high phenotypic plasticity, both within and between host species.
Haemosporida are vector-borne parasitic protozoa known to be present in birds of most avian orders. However, despite their perceived diversity using DNA barcode approaches, describing and delimiting species is challenging, particularly for those parasites found in non-passerine birds. In this study, we describe Haemoproteus trarotraro n. sp., a species found in two Falconiform hosts, the Crested Caracara (Caracara plancus plancus, type host) and the Yellow-headed Caracara (Daptrius chimachima chimachima), both sampled in Brazil at a wildlife rehabilitation center using microscopy and molecular tools. Haemoproteus trarotraro n. sp. is distinguished from the two other haemoproteid species described in Falconiformes, H. brachiatus and H. tinnunculi , by the absence of gametocytes that fully encircle the host-cell nucleus, and by the presence of numerous small vacuoles scattered throughout the cytoplasm of macrogametocytes. Both the partial cytb gene and the mtDNA genome for this new species are reported. The sequencing of the cytb barcode fragment revealed that H. trarotraro n. sp. reported here corresponds to a Haemoproteus sp. haplotype (GenBank Accession (AF465594) lineage POLPLA01 in Malavi) previously reported from Caracara plancus cheriway in Florida, USA. Although it diverges by only  2% at the cytb level from H. tinnunculi and H. brachiatus, H. trarotraro n. sp. is not a sister lineage to these taxa. Instead, phylogenetic analyses place it within a distinct but closely related, well-supported clade comprising lineages infecting American Kestrels (Falco sparverius). This study contributes, through an integrative taxonomic approach, to the ongoing discussion about species delimitation within the order Haemosporida.
The current outbreak of highly pathogenic avian influenza (HPAI) is the largest panzootic ever recorded, impacting millions of wild and domestic birds worldwide. Waterbirds including waterfowl, seabirds and shorebirds represent natural reservoirs for the virus and have facilitated its spread throughout major flyways. Peregrine falcons are obligate bird consumers and are considered HPAIV-susceptible. We investigated the influence of waterbird exposure on breeding peregrine falcons within the mid-Atlantic region of North America by comparing inland and coastal subpopulations. We monitored individually marked adults (N = 205) and breeding territories (N = 79) to estimate spatial and temporal patterns of adult survival, recruitment age and territory occupancy (2016-2025). Adult survival was comparable between inland and coastal locations prior to the arrival (date of first detection) of HPAIV but fell from 0.82 to 0.25 in coastal areas in the two years following arrival. Recruitment of juvenile-plumaged (second calendar year) birds was uncommon in both areas prior to the arrival of HPAIV but increased from 3.5% to 21.0% following arrival, suggesting that the recruitment pool may be diminished by the demand generated by increased adult turnover. Territory occupancy was high and stable prior to the arrival of HPAIV but declined by more than 50% in coastal areas following arrival. Our results document changes in adult survival, territory occupancy and juvenile recruitment that coincide with the appearance of HPAIV on the Atlantic Coast and suggest that impacts may be diet-mediated.
Blast furnace sludge (BFS) is an iron and carbon rich by-product whose direct recycling to ironmaking routes is often constrained by its fine particle size and the presence of volatile/corrosive species. This study evaluates Falcon enhanced gravity separation as a physical pre-treatment for iron upgrading from fine-grained BFS and optimizes operating conditions using Response Surface Methodology (RSM) with a three-factor, three-level Box-Behnken design (solid ratio 10-30%, fluidization water pressure 0.5-1.5 psi, and G-force 50-250 G). Statistical modeling confirmed that fluidization water pressure and G-force dominate separation performance, producing a clear grade-recovery trade-off: increasing water pressure improves Fe grade but reduces recovery, whereas increasing G-force tends to increase recovery while diluting grade. The BFS feed contained 34.40% Fe, 18.44% C and 2.94% Zn, under multi response optimization, the validation test yielded a concentrate grading 51.42% Fe at 57.88% Fe recovery. A two-stage scavenging configuration increased overall recovery to 77.52% at 49.35% Fe in the final concentrate. The results demonstrate that an optimized, two-stage Falcon circuit can effectively upgrade BFS and reduce mass for downstream treatment, while noting that Zn remains concentrated in the Fe-rich stream and may require subsequent removal depending on the intended metallurgical route.
Reports of audible sonic booms along the south-central California coast during SpaceX Falcon 9 launch ascents prompted measurements in Ventura County during summer 2024. A total of 132 measurements were made over six launches, with 16-25 measurements per launch. The maximum overpressure measured was 1.90 psf (133 dB), but most measured booms had an overpressure below 0.5 psf and durations of several seconds. Two launches had appreciably lower overpressures and smaller terrestrial footprint, indicating that both meteorology and launch azimuth are important factors in terrestrial boom audibility. Agreement between this dataset and environmental assessment predictions was marginal.
Hernán Dario Cerón-Muñoz was not included as an author in the original publication [...].
The saker falcon (Falco cherrug) is an Endangered avian predator that faces multiple threats within its wintering habitats, particularly in Iran. In this study, we aimed to model habitat suitability across the species' wintering range, identify the key environmental predictors of suitable habitat, and assess the extent to which these habitats are covered by protected areas in Iran. We applied an ensemble habitat suitability modeling approach (Generalised Boosting Models (GBM), Support Vector Machines (SVM), Maximum Entropy modelling (Maxent), and Random Forest (RF)) to predict the distribution of suitable wintering habitats for the saker falcon across Iran. Our results indicate that the most suitable wintering habitats are concentrated in northeastern and southern Iran, especially around the Persian Gulf and the Gulf of Oman. Annual precipitation with 41% contribution, NDVI with 21% contribution and distance from shrublands with 15% contribution emerged as the most influential predictors of habitat suitability. We also found that suitable habitats for the saker falcon are poorly represented within the current protected areas network. The suitable but unprotected habitats identified in this study may play a crucial role in the conservation of the species and should therefore be considered for inclusion within the protected areas system. Overall, our findings provide important insights for the conservation of this Endangered falcon, benefiting not only its wintering populations in Iran but also its global populations.
Falcons (genus Falco) are a rapidly diversified bird clade, with 38 species evolving over the past ∼7.5 million years. Despite their ecological and evolutionary significance, high-quality genomic resources remain limited. Here, we present two chromosome-level genome assemblies for Falco peregrinus and Falco biarmicus, enabling detailed analyses of genome architecture, gene content, transposable elements (TEs), and structural variation. Using PacBio HiFi sequencing, we generated highly contiguous assemblies (N50: 60.76-77.83 Mb) with >97% BUSCO completeness. Comparative analyses with Gallus gallus and Falco rusticolus revealed strong synteny among Falco species, whereas extensive chromosomal rearrangements were observed in comparison with the more distantly related Gallus gallus. TEs account for 7.43-8.44% of the genomes, with LINEs (CR1) and DNA transposons (Mutator, CACTA) predominating. Contigs belonging to the W sex chromosome were identified based on their significantly higher TE content (30% or more) compared to autosomes and the Z chromosome. Gene prediction, informed by long-read RNA Iso-Seq, identified 18,638-19,858 genes per genome, aligning with prior falcon annotations. Importantly we recovered key immune and sensory gene families, including MHC class I/II, innate immune receptors, and 24-25 olfactory receptor genes. We detected 8,746 structural variants, over 40% of which involved TEs, underscoring their role in genome polymorphism. These assemblies provide a valuable resource for investigating avian chromosome evolution, TE dynamics, and species-specific adaptations. They also establish a foundation for comparative genomics, population genetics, and conservation efforts in falcons.
Large Language Models (LLMs) have demonstrated strong performance in medical question-answering tasks, highlighting their potential for clinical decision support and medical education. However, their effectiveness in subspecialty areas such as nephrology remains underexplored. In this study, we assess the performance of open-source LLMs in answering multiple-choice questions from the Nephrology Self-Assessment Program (NephSAP) to better understand their capabilities and limitations within this specialized clinical domain. We evaluated the performance of five open-source large language models (LLMs): PodGPT which a podcast-pretrained model focused on STEMM disciplines, Llama 3.2-11B, Mistral-7B-Instruct-v0.2, Falcon3-10B-Instruct, and Gemma-2-9B-it. Each model was tested on its ability to answer multiple-choice questions derived from the NephSAP. Model performance was quantified using accuracy, defined as the proportion of correctly answered questions. In addition, the quality of the models explanatory responses was assessed using several natural language processing (NLP) metrics: Bilingual Evaluation Understudy (BLEU), Word Error Rate (WER), cosine similarity, and Flesch-Kincaid Grade Level (FKGL). For qualitative analysis, three board-certified nephrologists reviewed 40 randomly selected model responses to identify factual and clinical reasoning errors, with performance summarized as average error ratios based on the proportion of error-associated words per response. Among the evaluated models, PodGPT achieved the highest accuracy (64.77%), whereas Llama showed the lowest performance with an accuracy of 45.08%. Qualitative analysis showed that PodGPT had the lowest factual error rate (0.017), while Llama and Falcon achieved the lowest reasoning error rates (0.038). This study highlights the importance of STEMM-based training to enhance the reasoning capabilities and reliability of LLMs in clinical contexts, supporting the development of more effective AI-driven decision-support tools in nephrology and other medical specialties.
Neotropical raptors are among the most threatened birds, facing increasing extinction risks due to habitat loss and human persecution. Despite their importance for ecosystem stability, basic data on their distribution, abundance, and genetic diversity remain scarce. To address these gaps, we assembled and annotated the mitochondrial genomes of nine high-priority raptors from the Neotropics, including the threatened Chaco Eagle (Buteogallus coronatus), Black-and-Chestnut Eagle (Spizaetus isidori), Rufous-tailed Hawk (Buteo ventralis), and Harpy Eagle (Harpia harpyja), as well as the Near Threatened Orange-breasted Falcon (Falco deiroleucus), Crested Eagle (Morphnus guianensis), Ornate Hawk-Eagle (Spizaetus ornatus), Plumbeous Hawk (Cryptoleucopteryx plumbea), and Solitary Eagle (Buteogallus solitarius). Mitogenome sizes ranged from 17,848 to 20,449 bp, with consistent gene content and a Control Region architecture common in Falconidae and Accipitridae. Phylogenetic analyses provided strong support for most relationships, highlighting the value of mitogenomic data for phylogeographic studies. We further designed metabarcoding primers for environmental DNA applications. Primers targeting the 12S rRNA gene and a mini-barcode for the Harpy Eagle's Control Region showed high resolution using short, conserved sequences ideal for combining degraded DNA with next-generation sequencing. Our study provides essential molecular tools for monitoring and protecting these ecologically vital yet threatened raptors across the Americas.
The work offers a methodological improvement of decision-making based on a Bipolar Complex Fuzzy Multi-Criteria Group Decision-Making (BCF-MCGDM) framework that incorporates Aczel-Alsina triangular norms and conorms (AATNs and AATCNs) to become more effective at modeling uncertainty and dual experts’ preferences. The framework incorporates the merits of bipolar complex fuzzy sets (BCFS) and flexible aggregation operators in order to be able to embrace both positive and negative expert judgments in ambiguity. The proposed BCF-AATN approach, unlike the traditional fuzzy and intuitionistic fuzzy approaches, is able to address interdependent criteria and contradictory judgments in a complicated decision-making setting. To demonstrate its applicability, the framework is applied to a real-life problem in basketball strategy evaluation, where a number of experts evaluate tactical and analytical criteria. The model recognizes the Eastern Falcons as the most strategical team and justifies its strength using the sensitivity and comparative analyses. Generally, the suggested BCF-MCGDM framework provides a generalizable, theoretically based, and methodologically better system of decision support that is applicable to sports and other fields, which share certain characteristics of uncertainty and consensus among a group. Being a fast-paced and competitive game, basketball demands that players and coaches make decisions in changing tactics, uncertain conditions, and varying expert opinions. The suggested framework integrates BCF logic with adaptable aggregation techniques to assess complicated strategic alternatives. When applied to real-world data, the findings placed the Eastern Falcons as the most strategically appropriate team, with high tactical flexibility, awareness of the game, and data-driven strategies. The Northern Titans came in second, due to good player combinations and game analysis, followed by the Thunder Hawks. The Southern Stallions had a balanced performance in all categories, and the Capital Warriors stood last, as they depended on strategy but failed to execute it at the right time. Altogether, the framework based on the BCFS emphasizes the latent strengths and weaknesses of team strategies and offers viable information to coaches and analysts to facilitate the preparation of games, in-match decisions, and long-term planning.
Management of hepatocellular carcinoma (HCC) poses unique challenges due to its development in the context of chronic liver disease and the availability of multiple treatment options. Although multidisciplinary team (MDT) management improves outcomes, universal MDT discussion is resource-intensive, underscoring the need for effective patient-stratification tools. We developed a novel large language model (LLM) framework, PHENO-RAG, that integrates contemporary HCC management guidelines with patient-specific clinical data. We retrospectively analysed 489 clinical reports from 424 patients treated at a tertiary referral centre between September 2020 and November 2024. Eight locally hosted LLMs were tested: Llama-3-8B/70B, GPT-oss-20B/120B, Qwen-3-8B/80B, and Falcon-7B/40B. Two ablation studies assessed clinical concept extraction (using REGEX, pure LLMs, and hybrid REGEX+LLM pipelines) and decision generation across six configurations (zero-shot/few-shot with unstructured vs. structured notes, with and without retrieval-augmented generation [RAG] using clinical guidelines). The primary outcome was exact-match accuracy against real-world clinical decisions for treatment allocation, clinical complexity, and recommendation for MDT discussion. GPT-oss-120B+REGEX achieved the best overall agreement (median F1 for categorical concepts 0.92 [95% CI 0.85-0.95]; median intraclass correlation coefficient for numerical parameters 0.93 [95% CI 0.85-0.94]). For decision support, accuracy increased with structured inputs, few-shot exemplars, and RAG across all models. Under the strongest configuration (few-shot+RAG on structured notes), GPT-oss-120B reached 86.5% exact match for treatment allocation, 88.6% for clinical complexity, and 66.9% for MDT recommendation; Llama-3-70B achieved 80.8%, 83.4%, and 63.0%, respectively. Performance in the baseline zero-shot, unstructured-note configuration was substantially lower. PHENO-RAG delivers accurate, guideline-concordant support for HCC treatment allocation and complexity grading from real-world notes, with performance driven less by model family alone than by hybrid extraction, input structuring, in-context examples, and evidence retrieval. MDT referral remains the hardest task - appropriate for prioritization rather than automation. Prospective, multi-site and multimodal validation is warranted. Clinical decisions in the management of hepatocellular carcinoma are complex and multiparametric, requiring resource-intensive multidisciplinary care and creating challenges for optimal treatment allocation across different healthcare settings. We developed PHENO-RAG, a large language model-based framework that combines patient phenotyping through automated clinical information extraction from real-world clinical notes with treatment decision support, based on international guidelines. Our framework demonstrated concordance of 86.5% with real-world clinical decisions for treatment allocation and 88.6% for clinical complexity assessment, suggesting potential to enhance decision consistency and quality of care. In clinical practice, this AI-assisted framework could help standardize hepatocellular carcinoma management workflows, support training of hepatology and oncology fellows, assist in quality assurance programs, and facilitate more systematic identification of complex cases requiring multidisciplinary consultation, particularly in resource-constrained settings.
Bio-inspired flow-sensing and actuation mechanisms offer a promising path for enhancing the stability of flapping-wing flying robots (FWFRs) operating in dynamic and noisy environments. This study introduces a bio-inspired sensing and actuation feather unit (SAFU) that mimics the covert feathers of falcons and serves simultaneously as a distributed flow sensor and an adaptive actuation element. Each electromechanical feather (EF) passively detects airflow disturbances through deflection and actively modulates its flaps through an embedded actuator, enabling real-time aerodynamic adaptation. A reduced-order bond-graph model capturing the coupled aero-electromechanical dynamics of the FWFR wing and SAFU is developed to provide a physics-based training environment for a proximal policy optimization (PPO) based reinforcement learning controller. Through closed-loop interaction with this environment, the PPO policy autonomously learns control actions that regulate feather displacement, reduce airflow-induced loads, and improve dynamic stability without predefined control laws. Simulation results show that the PPO-driven SAFU achieves fast, well-damped responses with rise times below 0.5 s, settling times under 1.4 s, near-zero steady-state error across varying gust conditions and up to 50% alleviation of airflow-induced disturbance effects. Overall, this work highlights the potential of bio-inspired sensing-actuation architectures, combined with reinforcement learning, to serve as a promising solution for future flapping-wing drone designs, enabling enhanced resilience, autonomous flow adaptation, and intelligent aerodynamic control during operations in gusts.