Browse Articles

Backbone 3.0: An R package for extracting network backbones

The backbone package for R implements models for extracting the backbone – a sparse unweighted network containing only the most ‘important’ edges – from a weighted or unweighted network, where different models adopt different perspectives on what makes an edge important. The use of network backbones simplifies analysis and visualization when the original network is weighted or dense. This paper introduces and demonstrates the use of the backbone package for R to extract network backbones. After providing an overview of backbone ’s workflow and structure, I describe several backbone extraction models, illustrating backbone extraction in a series of toy examples. I then present a complete empirical case study using bill sponsorship data from the 108 th U.S. Senate. I conclude with some recommendations for backbone extraction, and an agenda for planned extensions to backbone .

Weaker and more frequent Mediterranean sea breezes in a warming climate

Abstract The Mediterranean, a climate change hotspot, has experienced accelerated warming, which is expected to influence sea breezes (SB) speed and occurrence. The extent to which these local winds have changed remains unclear due to the scarcity of multidecadal observations. Here we present a regional-scale assessment of SB across the Western Mediterranean, based on an unprecedented 41-year homogenized observational database spanning multiple countries and stations. Since the 1980s, SB have become weaker while more frequent. We link these changes to the interplay of thermodynamic and dynamic mechanisms driven by regional warming. The warming-induced weakening of SB speeds is further amplified during heatwaves, with speed reductions of up to 10%. In contrast, their winter occurrence has risen by roughly 10% per decade, consistent with a higher frequency of anticyclonic conditions. These findings offer a new understanding of SB response to the warming of the Mediterranean, with key implications for climate-related risks such as heat extremes and air quality.

Femtojoule optical Kerr switching with milliwatt-peak-power in silicon-organic hybrid nanocavity

Abstract All-optical Kerr switches enable ultrafast switching speeds, essential for next-generation communication and computing. However, their practical deployment is hindered by the intrinsically weak optical nonlinearity, which necessitates high switching energies. Although ultrashort pulses with Watt-level peak powers and extremely low duty cycles can alleviate this requirement, they are incompatible with real-world systems that demand high-duty-cycle data. In this work, we overcome this long-standing challenge by experimentally demonstrating an optical Kerr switch compatible with mainstream communication signals, while maintaining femtojoule-level switching energy. The breakthrough is achieved through a combination of exceptionally tight optical field confinement and a high-nonlinearity polymer material in a silicon-organic hybrid slot nanobeam cavity. This design reduces the required peak power to milliwatt-level, approximately two orders of magnitude lower than previous Kerr switches, enabling error-free switching of a 33%-duty-cycle 40-Gbit/s signal with 7.5 mW peak power and 63 fJ/bit switching energy. These achievements pave the way for chip-scale all-optical switches for ultrafast photonic signal processing.

Inhibition of purine nucleoside and nucleobase transporters by tyrosine kinase inhibitors

Tyrosine kinase inhibitors (TKI) are often used in combination with other chemotherapeutic nucleoside/nucleobase analogues, such as gemcitabine and 6-mercaptopurine, in the treatment of various cancers. Past studies have shown that several TKI inhibit the cellular uptake of nucleoside analogues by the equilibrative nucleoside transporter subtype 1 (ENT1), and suggest that TKI may also inhibit other related nucleoside and nucleobase transporters such as ENT2 and the equilibrative nucleobase transporter (ENBT1). To assess this possibility in a controlled manner, we have compared the ability of a series of TKI to inhibit each of these transporters in HEK293 cells that have been genetically modified to express either ENT1, ENT2 or ENBT1 in isolation, and on ENBT1 natively expressed in the chronic myeloid leukemia cell line K562. All TKI tested inhibited ENT1 and ENT2 with K i values ranging from 1 to 30 µM, typically with higher affinities for ENT1 than for ENT2. Gefitinib, which was one of the most effective inhibitors of ENT1, also inhibited ENBT1 with a similar affinity. The loss or gain of these transporters had no impact on the ability of gefitinib to directly affect cell viability, indicating that they were unlikely to be involved in the cellular uptake of the TKI. These data suggest that TKI inhibit multiple purine transporters, likely via interactions with their common purine ring binding domain. However, interactions between TKI and other nucleoside/nucleobase analogue drugs that are substrates for these systems are not likely to be a significant concern at the doses commonly used therapeutically.

Seismic attenuation transients reveal progressive crustal modification before and during the 2023 Türkiye earthquakes

Abstract We present the first spatiotemporally resolved separation of scattering and intrinsic seismic attenuation across the East Anatolian Fault Zone, capturing their evolution before and during the 2023 Türkiye earthquake doublet. Our results reveal a fault-segment–dependent decoupling between scattering and intrinsic absorption, indicating that the two components are governed by distinct physical processes rather than a single attenuation mechanism. Scattering attenuation intensified along the Çardak-Sürgü rupture, delineating spatial variations associated with strong rheological contrasts, whereas intrinsic attenuation exhibits broader temporal variability. Time-series analysis identifies measurable attenuation changes, reaching up to ~ 40%, associated not only with the mainshocks but also with moderate pre-sequence seismicity. These observations indicate that seismic attenuation records a temporal fingerprint of evolving crustal conditions. Overall, our findings demonstrate that resolving attenuation components provides diagnostic capability by capturing crustal modifications beyond the general post-earthquake damage observation.

Correlated quantum shift vector of particle-hole excitations

Abstract Excitons are a prime example of how electron interactions affect optical response and excitation. For example, electron-hole interactions produce a bound excitonic spectrum. Here we show that, beyond its spectra, the bound nature of an exciton’s electron-hole pair produces a correlated quantum geometry: excitonic excitations possess a quantum shift vector that is independent of light polarization. We find this counterintuitive behavior has dramatic consequences for geometric response: e.g., in noncentrosymmetric but non-polar materials, vertical excitonic transitions possess vanishing shift vector zeroing their shift photocurrent; this contrasts with finite and strongly light polarization dependent shift vectors for non-interacting delocalized particle-hole excitations. This dichotomy makes shift vector a sharp diagnostic of the pair localization properties of particle-hole excitations and demonstrates the non-perturbative effects of electron interactions in excited state quantum geometric response.

Enhanced nitrogen removal via simultaneous nitrification and denitrification by a newly isolated strain Enterobacter cloacae GW6 from estuarine sediment

Excessive nitrogen has been discharged into natural environments due to anthropogenic activities, leading to numerous negative impacts on natural ecosystems worldwide and becoming a global environmental issue. Therefore, effective removal of nitrogenous contaminants from natural ecosystems is crucial for the protection of both ecological sustainability and human interests. In this study, a strain of Enterobacter cloacae GW6 was newly isolated from natural estuarine sediment. This strain could utilize ammonium, nitrite and nitrate, with removal efficiencies of 99.60%, 93.73% and 98.79%, and the maximum removal rates of 39.96, 40.39 and 44.65 mg/L/h, respectively. Importantly, Strain GW6 showed excellent simultaneous heterotrophic nitrification and aerobic denitrification (HN-AD) capability with the maximum TN removal rate of 35.27 mg/L/h, which is greater than those of many other HN-AD bacteria reported previously, indicating that Strain GW6 was suitable for the efficient nitrogen removal in aquatic ecosystems with high ammonium, nitrite and nitrate concentrations. The detection of functional genes amoA , hao , napA , nirK and nosZ from Strain GW6 further confirmed its HN-AD capability. Strain GW6 exhibited strong nitrogen removal capability across a wide range of environmental conditions, which makes it a promising candidate as a HN-AD bacterium for the effective biological treatment of nitrogen contamination and eutrophication in natural aquatic ecosystems, thereby contributing to better environmental protection.

Phase Boundary Engineering of Co ₂ P‐CoP Branched Nanoparticles Enhances Cobalt Oxidation for Oxygen Evolution Electrocatalysis

ABSTRACT The boundaries between two different crystal phases contain atoms with unique electronic structures and coordination numbers that can significantly influence catalytic performance. Cobalt phosphide adopts Co 2 P and CoP crystal phases, and both are active for oxygen evolution reaction (OER), which offers the opportunity to improve catalytic activity through the creation of phase boundaries. Here we show that mixed‐phase Co 2 P‐CoP branched nanoparticles enriched with boundaries between the Co 2 P and CoP phases can be synthesized by controlled phosphidation of Co branched nanoparticles. We found that the slow transformation from Co 2 P to CoP is key to achieving Co 2 P‐CoP phase boundaries. These nanoparticles exhibit excellent OER performance with an overpotential of 240 mV that is 81 mV lower than that of a commercial RuO 2 standard, and is >3.5 times more active than the Co 2 P and CoP pure‐phase counterparts. Density functional theory calculations reveal that there is a partially positive charge stabilized on the Co atoms at the crystal phase boundaries that leads to enhanced OER activity. These results highlight the effectiveness of utilizing crystal phase boundaries in nanomaterials as a strategy for enhancing catalytic performance.

Automated Global Positioning Layout (GPL) for accuracy assessment in CAD-CAM mandibular reconstruction – method introduction

Abstract Assessing accuracy in CAD-CAM mandibular reconstruction poses significant challenges but is essential for ensuring reliable outcomes. Existing methods are often operator-dependent, lacking repeatability and reproducibility. This study introduces the Global Positioning Layout (GPL) method, an accuracy assessment technique integrated into the reconstruction protocol based on CAD-CAM and additive printing technology. We describe its methodology and present its implementation through an automated workflow. Key principles of accuracy assessment were identified and structured as Requirements, Data input, Data reference system, and Data output. The necessary 3D virtual models were defined: planned mandible, reference mandible, patient-specific implant (PSI), postoperative mandible, and postoperative PSI. A unique coordinate system (GPL-RS) was built on the reference mandible. Three Roto-Translational Matrices (RTMs) were applied to measure movements and deviations between the designed and postoperative models to assess reconstruction accuracy. Five clinical cases with different operational diseases and defects were analysed, comparing spatial deviations between manual and automated methods. The GPL method represents a promising advancement in assessing the accuracy of CAD-CAM reconstructions, providing valuable insights that can improve surgical outcomes.

Arid3b suppresses CD8 + T cell infiltration and function in microsatellite-stable colorectal cancer via Runx3

Abstract Microsatellite-stable/proficient mismatch repair (MSS/pMMR) colorectal cancer (CRC) is characterized by a cold tumor microenvironment, with limited CD8⁺ T cell infiltration and poor responsiveness to immune checkpoint inhibitors (ICIs). Here, using an in vivo CRISPR/Cas9 screen in a CMT93 cell-derived murine tumor model, we identify Arid3b as a key negative regulator of CD8⁺ T cell infiltration and antitumor activity. Genetic ablation of Arid3b in CD8⁺ T cells significantly enhances their intratumoral accumulation and promotes robust tumor control. Mechanistically, Arid3b deficiency upregulates Runx3 , driving a tissue-resident memory-like phenotype and effector function. Notably, the benefits conferred by Arid3b deficiency are abrogated upon Runx3 deletion, indicating a RUNX3-dependent mechanism. Together, targeting ARID3B could offer a promising strategy to reshape the tumor microenvironment and sensitize MSS CRC to immunotherapy.

Multidimensional performance characteristics of youth academy and club soccer players

Talent identification and development (TID) in soccer is complex. While physical performance and skeletal age (as an estimate of biological maturation) are well-established in TID research, the combined role of psychological skills and cognitive function remains underexplored. The aim of this study was to examine multidimensional performance characteristics among U14 players by comparing match-selected academy players (n = 20), non-match selected academy players (n = 14), and club players (n = 22). In total 56 players performed 40 m sprint test, Yo-Yo Intermittent Recovery Test Level 1 (IR1), and completed validated self-reported mental skills and motivation questionnaires. Executive functions were assessed using tests from Cambridge Neuropsychological Test Automated Battery (CANTAB) and skeletal age was assessed from left-hand X-ray images. The results showed that match-selected academy players were significantly taller, had higher weight and bone age compared to both club players and non-match-selected academy players. Match-selected academy players had significantly better physical performance and reported stronger mental skills, compared to club players. Club players reported higher levels of amotivation and external regulation compared to both match-selected and non-match-selected academy players, reflecting a lessself-determined motivational profile. Match-selected academy players demonstrated better decision-making, characterized by a significantly lower tendency to take risks to avoid delay and a greater ability to rationally adjust risk compared to club players. No further group differences were found for the executive function tests between groups.

Molecular docking, DFT analysis, pharmacokinetic profiling and MD simulation of Ilex aquifolium L. flavonoids as potential α-glucosidase predicted inhibitors for diabetes

Abstract The growing diabetes trend worldwide demands the identification of new α-glucosidase inhibitors for delaying postprandial hyperglycemia. However, the use of existing drugs is often limited by adverse effects, which has increased interest in natural products as safer alternative sources of α-glucosidase inhibitors. I. aquifolium was selected based on reports of flavonoid-rich phytochemistry, leading to the hypothesis that its major flavonoids could act as potential α-glucosidase inhibitors for the management of diabetes. Herein, the virtual α-glucosidase inhibitory potential of eight flavonoid-based phytochemicals of I. aquifolium extracts was investigated through an integrated in silico workflow, which includes molecular docking, DFT analysis, and ADMET profiling. Docking analysis identified IA- 2 (Rutin) and IA- 5 (Kaempferol-3-O-rhamnoglucoside) as the most potent ligands with binding affinities of − 9.59 kcal/mol and − 9.18 kcal/mol, respectively, closely approaching the standard acarbose (–10.96 kcal/mol). The re-docking RMSD of 2.1375 Å fell within the acceptable validation benchmark, supporting the reliability of the docking protocol. The second and third positions were occupied by IA- 1 (Quercetin) and IA- 3 (Quercetin-3-O-hexoside) as potential inhibitors, as they formed hydrogen bonds with hydrophobic contacts. DFT study suggested that all the flavonoids exhibited moderate to high dipole moments (4.88–7.36 Debye) along with the high HOMO–LUMO gaps. Additionally, IA- 7 (Apigenin) showed the highest electrophilicity (3.313) among all the phytochemicals. The pharmacokinetic evaluation of ligands by SwissADME and PkCSM suggested IA- 1 as the most potential hit, and it occupied the main drug-likeness parameters with a bioavailability score of 0.55, whereas low bioavailability scores (0.17) for IA- 2 and IA- 5 indicated their poor permeability through the membrane. The potential toxicity of the ligands through SwissADME predicted that all the ligands have renal and respiratory issues, but IA- 1 showed the highest potential of acute dermal toxicity. MD simulations supported the docking results by showing protein backbone RMSD stabilization at 2.0–2.4 Å and revealing that IA-2 and IA-5 form the most dynamically stable complexes (lowest ligand RMSD, compact Rg, persistent H bonds), while IA-3 and IA-1 display greater ligand flexibility, thereby prioritizing these compounds for experimental validation in a hypothesis-generating context. However, these computational findings do not replace experimental validation and should be interpreted as hypothesis-generating predictions. Overall, the study provides a hypothesis-generating framework, prioritizing IA-2 and IA-5 as potential α-glucosidase inhibitory candidates, with IA-1 emerging as a top-ranked ligand through pharmacokinetic parameters, required further experimental validation.

Orbital-resolved imaging of coherent femtosecond exciton dynamics in coupled molecules

Abstract Optical excitation and control of excitonic wavepackets in organic molecules is the basis to energy conversion processes. To gain insights into such processes, it is essential to establish the relationship between the coherence timescales of excitons with the local electronic distribution in the molecules, as well as the influence of intermolecular interactions on exciton dynamics. Here, we demonstrate orbital-resolved imaging of optically induced coherent exciton dynamics in single copper naphthalocyanine (CuNc) molecules, and selective coherent excitation of dark and bright triplet excitons in coupled molecular dimers. Ultrafast photon-induced tunneling current enabled atomic-scale imaging and control of the excitons in resonantly excited molecules by employing excitonic wavepacket interferometry. Our results reveal an ultrafast exciton coherence time of ~ 70 fs in a single molecule, which decreases for the triplet excitons in interacting molecules.

D-dimer and lower limb ultrasound as prognostic factors for recurrent deep venous thrombosis and pulmonary embolism: A systematic review and meta-analysis

Venous thromboembolic disease is a chronic, recurrent condition. The optimal duration of anticoagulation therapy remains uncertain. We aim to evaluate D-dimer and lower limb ultrasonography as prognostic tools for the recurrence of venous thromboembolism. A search was conducted on May 28, 2022, in the Medline, Embase, and Cochrane databases. Inclusion criteria encompassed cohort studies, case-control studies, and clinical trials. Two reviewers independently screened all records and analyzed studies for inclusion/exclusion criteria, as well as risk of bias, using a structured framework (PROSPERO ID: CRD42022341082). The initial search yielded 4652 titles and abstracts. After removing 777 duplicates and reviewing 3875 titles and abstracts, 48 articles providing information on D-dimer and/or lower limb ultrasonography as prognostic factors were finally included. Very low-certainty evidence suggests that both residual vein thrombosis (RVT) at anticoagulation discontinuation and a positive post-treatment D-dimer are significantly associated with an increased risk of recurrent venous thromboembolism (VTE). For RVT, the pooled analysis showed a two-fold increase in risk (OR 2.00, 95% CI 1.02 to 3.94; I² = 85.7) while for positive D-dimer, the risk was similarly elevated (OR, 2.48; 95% CI, 1.85–3.33; I² = 56.0). In conclusion, very low-quality evidence suggests that abnormal D-dimer and RVT are associated with recurrent VTE; however, this association is inconsistent due to significant heterogeneity and wide prediction intervals. These biomarkers should be interpreted with caution as isolated predictors in clinical practice.

Performance and statistical evaluation of animal dung–driven microbial fuel cells for simultaneous bio electrochemical conversion and bioremediation of toxic metal-contaminated wastewater

Computational engineering of the polyester hydrolase PHL7 for efficient poly(ethylene terephthalate) degradation in biocatalytic recycling processes

Abstract Polyethylene terephthalate (PET) plastic waste causes serious environmental pollution due to insufficient recycling rates. Enzymatic PET depolymerization offers a sustainable recycling strategy, but limited stability and activity of current PET-degrading enzymes restrict practical implementation. Here, we engineer Polyester Hydrolase Leipzig 7 (PHL7), a PET hydrolase from a compost metagenome, to enhance its stability and catalytic performance under recycling-relevant conditions. Using Rosetta PROSS-based computational design combined with rational mutagenesis, we introduce up to 24 mutations, generating variants with melting temperatures of 88-95 °C and over 110-fold higher activity in 0.1 M phosphate buffer compared to the parent enzyme. Benchmarking shows that the best variants (R4M6, R4M9, and R4M10) match or exceed the performance of established engineered PET hydrolases, including ICCG and LCC-A2, and approach that of TurboPETase across multiple conditions. Under high substrate loadings, the PHL7-R4 variants degrade 75-78% of 10% (w/w) PET within 24 h at 65 °C, outperforming ICCG, while an optimized variant R4M10-H185Y achieves up to 84% degradation of 20% (w/w) PET. X-ray structure determination and molecular dynamics simulations reveal key stabilizing and activity enhancing mechanisms. These engineered PHL7 variants represent robust biocatalysts for scalable enzymatic PET recycling.

Antibiotic dispensing practices and antimicrobial stewardship gaps in community pharmacies in Kakamega County, Kenya

Background Antimicrobial resistance (AMR) is a major global health threat, with sub-Saharan Africa bearing a disproportionate burden. Community-level antibiotic dispensing practices remain poorly described in Kenya outside Nairobi. Methods A total of 504 antibiotic dispensing events were recorded across 22 community pharmacies in Kakamega County, western Kenya, between 3rd and 22nd August 2025. Data collected included dispensing source (over-the-counter [OTC] versus prescription), clinical indication, antibiotics dispensed, course completion, and self-reported repeat antibiotic use within the preceding month. Descriptive analyses were performed, and χ² tests were used to examine associations between dispensing source and selected non-antibiotic dispensing characteristics. Results Of the 504 dispensing events, 224 (44.4%) involved OTC dispensing and 278 (55.2%) were prescription-based. The most frequent indications for antibiotic dispensing were upper respiratory tract infections (URTI; n = 156, 31.0%), lower respiratory tract infections (LRTI; n = 95, 18.8%), gastrointestinal infections (n = 65, 12.9%), and skin or soft-tissue infections (n = 55, 10.9%). Across all events, amoxicillin, azithromycin, and metronidazole were the most frequently dispensed antibiotics, with cephalosporins and other broad-spectrum agents used across several indications. Partial antibiotic courses were supplied in 33 (6.5%) dispensing events, most commonly due to financial constraints (15/33, 45.5%). Self-reported antibiotic use within the preceding month occurred in 156 (31.0%) cases. Conclusions OTC antibiotic access remains widespread in Kakamega County, with substantial use of broad-spectrum agents across multiple clinical indications. Financial barriers contribute to incomplete antibiotic courses. These findings highlight the importance of incorporating community pharmacy dispensing data into county-level antimicrobial stewardship programmes and informing national strategies to optimise antibiotic use.

Miniature Soft Robot With Magnetically Reprogrammable Surgical Functions

ABSTRACT Magnetic miniature robots are untethered actuators, which can prospectively make existing minimally invasive surgery considerably safer and painless, and enable unprecedented treatments because they are much smaller and dexterous than existing surgical robots. However, such actuators are restricted to possessing at most two on‐board functionalities, having limited five degrees‐of‐freedom (DOF) locomotion, or are only operational under specialized environments where actuation from strong external magnets must be at very close proximity (<4 cm away). Here we present smart magnetic composites with deliberately engineered symmetry and heterogeneous material properties. Leveraging these composites, we construct a millimeter‐scale soft robot whose magnetization profile can be reprogrammed to dispense drugs, cut through biological tissues, grip and store biological samples, and heat remotely. By possessing full six‐DOF motions, including the sixth‐DOF rotation about its net magnetic moment, our soft robot can also roll and two‐anchor crawl across challenging unstructured environments, which are impassable by its five‐DOF counterparts. Because the applied magnetic fields are relatively uniform and weak (10–30 mT, 0.21–0.31 T/m for actuation; 60–65 mT for reprogramming), they can theoretically penetrate human tissues safely and enable actuation beyond 5 cm.

Network clustering algorithms and preprocessing pipelines for robust cell type identification in single-cell RNA sequencing data

Abstract Single cell RNA-seq (scRNA-seq) technologies provide unprecedented resolution representing transcriptomics at the level of single cell. One of the biggest challenges in scRNA-seq data analysis is the cell type annotation, which is usually inferred by cell separation approaches. In-silico algorithms that accurately identify individual cell types in ongoing single-cell sequencing studies are crucial for unlocking cellular heterogeneity and understanding the biological basis of diseases. In this study, we focus on robustly identifying cell types in single-cell RNA sequencing data; we conduct a comparative analysis using methods established in biology, like Seurat, Leiden, and WGCNA, as well as network-based methods Infomap, statistical inference via Stochastic Block Models (SBM), and single-cell Graph Neural Networks (scGNN). We also analyze preprocessing pipelines to identify and optimize key components in the process, explicitly considering their role in mitigating inherent data noise and potential batch effects for robust cell type identification. Leveraging three independent datasets, PBMC, ROSMAP, and MOp, we employ clustering algorithms on cell-cell networks derived from gene expression data. Our findings reveal that clusters identified by multiresolution Infomap and Leiden show a closer alignment, with Infomap standing out as a particularly effective approach. Infomap notably offers valuable insights for the precise characterization of cellular landscapes related to neurodegeneration and immunology in scRNA-seq.

A high dimensionality approach reveals immunopathogenic responses driving severe pediatric acute respiratory distress syndrome

Abstract Mechanisms underlying paediatric acute respiratory distress syndrome (PARDS) remain poorly understood, limiting advances in diagnosis and treatment. To address this, we conduct a high-dimensional, multi-omics analysis of paired pulmonary and blood samples from children with PARDS and age-matched controls. Our approach includes transcriptomics, proteomics, flow and mass cytometry, and single-cell RNA sequencing, with further validation using cytokine assays and in vitro models. Severe PARDS is characterised by three convergent immune abnormalities; Pulmonary CD8 + T cells display an interferon-driven cytotoxic profile, with exhaustion and apoptosis genes; Pulmonary T cells and myeloid cells exhibit strong interferon-stimulated gene expression; Distinct macrophage subsets show high interferon but suppressed IL-1 pathway genes, associated with impaired leukocyte chemotaxis, phagocytosis, and M1-polarization. This is mirrored by reduced pulmonary IL-1α/β and elevated IFN-γ. The systemic IL-1 signature is similarly dampened, while interferon responses are compartmentalised to the lung. Using an in vitro model, IFN-γ priming is shown to suppress TLR7-induced IL-1 β production through transcriptional inhibition of downstream inflammatory pathways, recapitulating the immune signature observed in patients. Our results reveal interferon-driven immune dysregulation and IL-1 suppression as central features of severe PARDS, highlighting parallels and differences from adult ARDS and underscoring the need for paediatric-specific therapeutic strategies.