Efficient, cell type-selective delivery of genetic payloads remains a central challenge in the development of gene and cell therapies. Lipid nanoparticles (LNPs) offer a versatile delivery platform, but their optimization is hindered by reliance on brute-force screening methods that are laborious, resource-intensive, and focus on single targets. Here, we present FALCON (Framework for Active Learning-driven Compositional Optimization of Nanoparticles), a closed-loop pipeline that leverages iterative screening, surrogate modeling, and multi-objective optimization to accelerate LNP compositional design. In B cell-targeted validation experiments, FALCON-optimized LNPs achieved a 1.8-fold increase in splenic B cell transfection in vivo compared with reference compositions. When optimized for selectivity, FALCON LNPs displayed an 84-fold improvement in selective transfection of splenic B cells over off-target liver populations and enabled spleen-tropic behavior across factorial panels of varying ionizable and helper lipid chemistries. In vaccine studies, these LNPs induced higher IgG2c antibody titers and a more Th1-biased immune profile. FALCON was also deployed to optimize LNPs for myeloid cell-selective delivery, achieving enhanced in vivo selectivity following systemic administration both across and within spleen and liver compartments. Our results establish FALCON as a useful tool for data-driven design of LNP compositions for precision gene delivery.
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.
In birds, the ability to produce vocalizations learned from conspecifics (vocal production learning, VPL) is well-characterized from an acoustic and neurological perspective, but the evolution of sound source morphology in clades with vocal learners remains poorly known. The syrinx, the vocal organ of birds, shows increased complexity in the three known avian clades with this trait-passerines, parrots, and hummingbirds-but evolutionary patterns preceding, and associated with, inferred acquisitions of VPL have never been studied in a phylogenetic context. Here we redescribe the syrinx of eufalconimorphs-passerines, parrots, and the non-learning falconiforms-and investigate evolutionary shifts in its morphology using phylogenetic comparative methods. Syringeal muscle insertions on, or closely adjacent to, sound sources are estimated to be ancestral to eufalconimorphs, while complex intrinsic syringeal muscles in passerines and parrots are recovered as two independent acquisitions. Passerines have several unique syringeal traits, including enlarged labia, optimized to occur prior to oscine VPL acquisition. We hypothesize that these traits may have facilitated later independent acquisitions of precise motor control of sound sources associated with complex vocalizations. We discuss other traits that may be associated with vocal production differences among clades, including variation in upper vocal tract anatomy and life-history characters.
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.
Diabetic Retinopathy (DR) is a prominent results of diabetes mellitus that causes abnormalities lesions in retina. If not identified at early, it may progress to complete loss of vision. Unfortunately, DR is an irreversible, and treatment only sustains existing vision. Timely detection and accurate treatment of DR can considerably decrease the chance of blindness. Manual diagnosis of DR in retinal fundus images (RFIs) by ophthalmologist is time consuming, costly and laborious tasks with a higher risk of misdiagnosis. Recently, Deep learning (DL) has gained popularity and shown remarkable performance particularly in medical image analysis and classification. Convolutional neural networks (CNNs) are increasingly being used as a DL approach in medical image analysis, and they are very efficient. This manuscript offers the design of Falcon Optimizer with Ensemble of Deep Learning Algorithm Assisted Diabetic Retinopathy Diagnosis Model (FOEDLA-DRDM)  system on RFIs. The FOEDLA-DRDM system employs a Wiener filtering (WF) based preprocessing approach to eliminate noise from images. Following this, FOEDLA-DRDM system leverages the SE-DenseNet method to generate the feature vectors. For DR recognition FOEDLA-DRDM system applies an ensemble approach that combines - AutoEncoder, long short-term memory (LSTM), and deep belief network (DBN). Finally, Falcon Optimizer (FO) adjusts the hyperparameter values of the ensemble approach, giving rise to classification efficiency. The FOEDLA-DRDM system is validated by simulating it on a Kaggle DR dataset, with results being measured according to various criteria. The simulation findings showcase the effectiveness of the FOEDLA-DRDM system in diagnosis of DR.
Digital vaginal examination is the technique that midwives use to assess the stage of labour, but its learning process is based on theory, mannequins and mostly practice with pregnant women, which often discomforts those women and does not assure learning all kinds of situations that midwives can face. Virtual reality can greatly help in learning this technique, but current virtual medical simulators are based on consumer headsets and controllers that do not provide the required haptic feedback. To develop a realistic, but also affordable, virtual reality obstetric simulator for training the palpation technique on different cervix's conditions and foetal head orientations. Specifically, this work is focused on the examination that midwives perform to find the fontanelles in the foetal head and thus determine its orientation, which is key to anticipating problems during labour. This development was carried out using the TRES-D methodology, which addresses 3D interfaces and content for virtual reality. The simulator is based on a low-cost haptic device, the Novint Falcon, which allows the user to feel the force produced when touching a virtual object. One midwife was directly involved in this process, and a guided evaluation with four other midwives was later conducted to check its realism and assess its potential. Preliminary tests gathered mostly positive feedback. Participants found the simulator useful and intuitive, and highlighted its tactile realism and detail, as they were able to assess the cervix and foetal configuration, touching the sutures and fontanelles as in a real examination. The developed obstetric simulator has the potential to become a useful tool in learning this examination technique. The main limitation to overcome in the future, and common to other-even more expensive- haptic devices, is having not only one point of contact but two.
暂无摘要（点击查看详情）
Falco biarmicus feldeggii has experienced increasing anthropogenic pressures over recent decades, resulting in regional population declines. Despite its critical conservation status, the species remains poorly characterized from a genetic standpoint. In this study, we assessed genetic differentiation between captive and wild Italian specimens of F. b. feldeggii to evaluate the genetic consequences of captive breeding. We also simulated alternative population reinforcement scenarios and developed a predictive model integrating demographic trends, population viability, and genetic outcomes based on actual genotypes and alternative mating systems. Our results showed a close genetic similarity between captive and wild specimens, supporting the use of the captive dataset as input for the simulation model. Simulation of various reinforcement scenarios highlighted that juvenile mortality had a stronger influence on the establishment of a long-term self-sustaining population. Furthermore, our model showed that while demographic parameters strongly shaped the trajectories of genetic diversity, selected mating had only a limited and short-term impact in our simulations providing a small contribution to stabilizing early-generation genetic diversity.
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.
Raptors, including the orders Accipitriformes (hawks and kites), Falconiformes (falcons and caracaras), Cathartiformes (New World vultures), and Strigiformes (owls), are found in small forest fragments, parks, vacant lots, outskirts, and open areas within the Metropolitan Region of São Paulo, in the state of São Paulo, Brazil. However, few studies have examined the infectious agents that infect them, particularly protozoa. This research reports on the seroprevalence, isolation, and genetic diversity of the zoonotic parasite Toxoplasma gondii in rescued raptors from two wildlife rehabilitation centres. These birds were fed live mice from a certified institution, as well as quails and insects from commercial establishments. A total of 151 raptor specimens was sampled, comprising five Cathartiformes, 30 Accipitriformes, 31 Falconiformes, and 85 Strigiformes, representing 19 species. Anti-T. gondii IgG antibodies were identified via the Modified Agglutination Test (MAT; cut-off ≥ 20). Bioassays in mice were performed to isolate T. gondii, and the genetic diversity of the isolates was examined via PCR-RFLP and microsatellite genotyping. Of the 151 birds, serum samples were collected from 150 specimens. MAT results showed that 62 birds (41.3%) across 14 species were seropositive, including 19 of 29 (65.5%) Accipitriformes, 19 of 31 (61.3%) Falconiformes, and 24 of 85 (28.2%) Strigiformes. Among the 128 bioassays conducted in mice, 27 (21.1%) T. gondii isolates were obtained from birds of nine species, including isolates from Rupornis magnirostris (7), Geranoaetus albicaudatus (2), and Elanus leucurus (1); Caracara plancus (9), Falco sparverius (3), and Falco femoralis (1); Asio clamator (2), Megascops choliba (1), and Asio stygius (1). PCR-RFLP genotyping identified 16 genotypes, and a mixed genotype, including genotypes #11 (Type BrII - 7 isolates), #19 (2), #21 (1), #22 (1), #33 (2), #51 (1), #69 (1), #111 (1), #162 (1), #175 (3) and five new genotypes designated #350, #351, #352, #353, and #354. Microsatellite analysis revealed 26 genotypes and a mixed genotype. Some rare alleles detected included 287 for TUB2, 246 for W35, 203 for TgM-A, 364 and 366 for B17, and 273 for MIV.1. Toxoplasma gondii is highly prevalent and genetically diverse among the wild raptors in the studied population. The same strains may circulate among wild raptors, domestic animals and humans.
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.
Aim: We aimed to provide a structured ex vivo protocol for cardiopulmonary micro-CT that combines gelatin-barium sulfate (gelatin-BaSO4) contrast medium with agar embedding in neonatal canine cardiopulmonary specimens. Materials and Methods: Heart-lung specimens from 23 puppies that died shortly after birth were collected, stored at -20 °C, and then slowly thawed prior to imaging. Before perfusion, body mass and heart-lung complex mass were recorded. Body mass ranged from 140 to 951 g, and heart-lung complex mass ranged from 1.2 to 51.2 g. The cranial and caudal venae cavae, the brachiocephalic trunk, and the left subclavian artery were ligated. A catheter was introduced into the thoracic aorta. Contrast was prepared by dissolving porcine gelatin in hot water and mixing with a commercial BaSO4 suspension. The mixture was maintained at a warm temperature to remain free-flowing and was delivered at low pressure until uniform opacification of the coronary and pulmonary arteries was observed. After in situ gelation, the organs were embedded in warm agar and sealed to limit motion and dehydration. Scans were performed on a benchtop system (120 kV, ~83 µA, ~1200 projections, ~2 s exposures; voxel ~40 µm). Reconstruction was performed in XMReconstructor, with post-processing in Falcon and RadiAnt. The reconstructed micro-CT datasets were reviewed anatomically by a medical cardiologist and a veterinary cardiologist, whereas vascular filling was evaluated semi-quantitatively by three observers with expertise in veterinary anatomy and cardiology. Results: In all specimens examined, the main coronary artery course was assessable. Conclusions: The gelatin-BaSO4 contrast medium combined with agar immobilization provides a simple, lead-free, and affordable approach for structured cardiopulmonary micro-CT in very small post-mortem specimens. In the examined specimens, the workflow provided visually consistent low-pressure vascular opacification without gross evidence of vessel rupture or motion-related acquisition failure under the conditions of this study. Practical mitigations included temperature/viscosity control, avoidance of phosphate buffers, container sealing, and minimization of particle aggregation, bubbles, and dehydration. The protocol may complement conventional autopsy in very small post-mortem specimens in similar ex vivo research settings.
Cryo-electron microscopy (cryo-EM) data acquisition is time-intensive given that a large amount of data is required to obtain a high-resolution reconstruction. Here, we eliminate camera-induced delay time by continuously recording during beam-image shift acquisition using a method called Continuous Recording Beam-Image Shift (CR-BIS). The utilization of CR-BIS with K3 and Falcon 4 direct electron detectors and conventional data acquisition conditions enables the acquisition of ~34,000 micrographs and ~1,000 tilt series per 24 h in single-particle analysis mode and cryo-electron tomography mode, respectively. Three-dimensional reconstructions of single-particle and tomographic datasets show that CR-BIS accelerates data collection and maintains data quality. CR-BIS is broadly applicable for efficient high-resolution cryo-EM since it can be implemented into existing acquisition software through scripting and it does not require hardware modification.
From falcons spotting prey to humans recognizing faces, the ability to rapidly process visual information depends on a foveated retinal organization that provides high-acuity central vision while preserving low-resolution peripheral vision. This organization is conserved along early visual pathways, yet remains under-explored in machine learning. Here, we examine the impact of embedding a foveated retinotopic transformation as a preprocessing layer on convolutional neural networks (CNNs) for image classification. By applying a log-polar mapping to off-the-shelf models and retraining them, we achieve comparable accuracy while improving robustness to scale and rotation. We demonstrate that this architecture is highly sensitive to shifts in the fixation point and that this sensitivity provides an effective proxy for defining saliency maps that facilitate object localization. Our results demonstrate that foveated retinotopy encodes prior geometric knowledge, providing a solution for visual searches and a meaningful classification robustness and localization trade-off. These findings provides a proof of concept in order to connect principles of biological vision with artificial networks, suggesting new, robust and efficient approaches for computer vision systems.
HPV primary testing is the preferred method for cervical cancer screening worldwide and enables vaginal self-sampling at home. Self-sampling performs comparably to clinician-collected samples and can improve participation, especially among non-attenders. In Finland, self-sampling has so far been piloted in the Helsinki region. This study evaluated the feasibility and acceptability of self-sampling in Pirkanmaa, providing a representative model of implementation outside the capital region. The self-sampling in cervical cancer routine screening (FALCON) study (NCT06931184) targeted women living in Pirkanmaa who did not respond to the initial cervical cancer screening invitation. In a reminder letter, they were offered an opt-in model self-sampling option. Participants ordered a self-sampling kit containing a FLOQSwab (Copan), instructions, and a prepaid return envelope. Samples were analyzed with the Roche Cobas 4800 HPV assay at Fimlab Laboratories. HPV-positive women were referred for cytology testing and managed according to national guidelines. Participants completed two online questionnaires: one at kit order and another 60 days later. Of 16 289 reminder letters, 329 kits were ordered, and 304 (1.9%) samples were returned. Participants (mean age 48 [range 29.9-65.8]) were mostly well-educated with previous screening history. Time-saving, reduced discomfort, and lower embarrassment were the main reasons for choosing home-sampling. HPV was detected in 6.9% of the samples, with one case of histological HSIL identified. Most participants found self-sampling easy and comfortable, and nearly all would recommend it to others. Despite the small number of participants in this first year, the study offers early evidence of the strong acceptability and practicality of self-sampling among women in Pirkanmaa, warranting further evaluation at a larger scale.
Long non-coding RNAs (lncRNAs) have emerged as crucial regulators in plant responses to abiotic stresses, including salinity, drought, heavy metals, and temperature fluctuations. The functional characterization of drought-responsive lncRNAs in Solanum lycopersicum remains incomplete and poorly understood. In this study, we performed transcriptome-wide identification and functional analysis of lncRNAs in two tomato cultivars, drought-tolerant variety Falcon, and drought-susceptible SC2121, when subjected to drought stress. Physiological parameters-such as shoot height, root length, soluble protein, and water content- were measured to evaluate drought-induced changes in the tomato varieties. In addition, five lncRNAs were selected for reverse transcription quantitative PCR (RT-qPCR) analysis. Using high-throughput RNA sequencing, we identified 269 drought-responsive lncRNAs, including 124 upregulated and 145 downregulated transcripts. Interestingly, in tomato, predicted lncRNA-mRNA interactions suggest that XR_003244833.1, XR_003247168.1, and XR_743350.3 may be associated with the sorbitol (polyol) pathway, the ABC-2 type transporter protein, and the U11/U12 small nuclear ribonucleoprotein involved in the spliceosome complex, respectively. Using high-throughput RNA sequencing, we identified 269 drought-responsive lncRNAs, including 124 upregulated and 145 downregulated transcripts. Interestingly, in tomato, predicted lncRNA-mRNA interactions suggest that XR_003244833.1, XR_003247168.1, and XR_743350.3 may be associated with the sorbitol (polyol) pathway, the ABC-2 type transporter protein, and the U11/U12 small nuclear ribonucleoprotein involved in the spliceosome complex, respectively. Accordingly, lncRNAs may be actively involved in the drought stress response and regulate key adaptive pathways through interactions with protein-coding genes. These findings provide new insights into the molecular mechanisms of drought tolerance in tomato and offer potential targets for the genetic improvement of stress-resilient cultivars.
Pulmonary large cell neuroendocrine carcinoma (LCNEC) is a high-grade neuroendocrine carcinoma with clinical and molecular heterogeneity distinct from small cell lung cancer (SCLC). This large retrospective series seeks to define clinical outcomes for patients with newly diagnosed metastatic LCNEC and whether aligning treatment based on molecular classification improves outcomes. Clinicopathological data included demographics, treatment history, and immunohistochemical and genomic results. Patients with suspected transformed LCNEC (tLCNEC), underwent curated pathology review. Patient outcomes were assessed using Kaplan-Meier methodology. 137 patients were included (100 with pure LCNEC and 37 with mixed histologies), of whom 15 had tLCNEC. For 115 cases with genomic data available, 63 (55%) were classified as NSCLC-like and 43 (37%) as SCLC-like. Median overall survival (OS) from start of first-line therapy was 10.0 months [95% CI 8.6-11.9] for patients with primary LCNEC and 16.3 mo (9.4-40) for tLCNEC. There was no statistical difference in OS between patients treated with NSCLC- or SCLC-regimens [mOS 12.1 (10.0-17.6) vs 9.4 mo (6.3-11.2)] or with the addition of immunotherapy to chemotherapy [median OS 9.9 mo (7.7-13.9) vs. 10.4 mo (8.6-14.1)]. Several patients with tLCNEC who received local therapies for oligometastatic disease had long-term disease control. Clinical outcomes for patients with LCNEC are poor, without clear association of benefit observed through application of molecularly-classified standard treatments. Based on these data, we propose guidelines for development of prospective trials of novel therapeutics aimed at addressing unmet needs of this patient population.
The standardisation of post-quantum cryptography (PQC) by NIST marks a critical transition away from classical public-key schemes towards quantum-resistant successors. As machine learning (ML) applications proliferate, the demand for efficient cryptographic primitives intensifies, requiring implementations that are simultaneously quantum-safe and resource-aware. Recent surveys have investigated the interplay between ML and PQC, with particular focus on ML-assisted parameter optimisation, privacy-preserving ML leveraging lattice-based cryptography, and neural-network implementations of quantum-resistant algorithms. Building on these findings, we propose QSafe-ML, a comprehensive four-stage framework that integrates hardware profiling, surrogate modelling via ML, constrained multi-objective optimisation, and continuous security validation to facilitate the tuning of PQC parameters and implementations. The framework targets NIST-standardised lattice-based schemes CRYSTALS-Kyber, CRYSTALS-Dilithium, Falcon, and NTRU across three heterogeneous hardware platforms. Experimental evaluation with n=30 repeated trials demonstrates mean latency reductions of 27.5-41.9% (95% CI ±1.1-1.7 pp), memory savings of 13.3-30.2%, and energy savings of 22.8-38.2% over NIST reference baselines, with all configurations maintaining ≥128-bit post-quantum security. An ablation study confirms that surrogate-guided search accounts for the dominant share of these gains. All code, data, and benchmark instructions are released at a public repository (available upon acceptance of this manuscript) to promote reproducibility in evaluating ML-assisted cryptographic systems.
Bats have unique immunological characteristics among vertebrates. Understanding how the immune system responds during the life cycle of bat populations is crucial for grasping how their defenses against infection vary during critical life events. We used the lesser long-nosed bat (Leptonycteris yerbabuenae) as a model to assess adaptive humoral immunity (total immunoglobulin G concentration, tIgG) at different periods during the year involving different ages, reproductive conditions and migratory habits (females migrate seasonally to give birth while males are resident). We also determined the relationship of tIgG with previously published values of plasma bacterial killing ability against Escherichia coli (BKA). tIgG did not vary among reproductive periods in males and females, although higher values were measured in non-reproductive males in 2019 than in 2020. In contrast, tIgG was lower in young individuals than in non-reproductive and lactating females. The tIgG had a positive relationship with BKA in non-reproductive females, but no relationship was observed in non-reproductive males and reproductive females. These results suggest that tIgG is resilient to physiological and ecological changes experienced by adults throughout their reproductive and migratory cycle. The low tIgG during early life suggests that young individuals have not developed an IgG pool comparable to that of adults. The positive relationship between tIgG and BKA in non-reproductive females suggests that E. coli clearance is enhanced by the presence of antibodies that facilitate its elimination.