PLoS ONE

BACKGROUND: the analysis of microbial communities through dna sequencing brings many challenges: the integration of different types of data with methods from ecology, genetics, phylogenetics, multivariate statistics, visualization and testing. With the increased breadth of experimental designs now being pursued, project-specific statistical analyses are often needed, and these analyses are often difficult (or impossible) for peer researchers to independently reproduce. The vast majority of the requisite tools for performing these analyses reproducibly are already implemented in R and its extensions (packages), but with limited support for high throughput microbiome census data. RESULTS: Here we describe a software project, phyloseq, dedicated to the object-oriented representation and analysis of microbiome census data in R. It supports importing data from a variety of common formats, as well as many analysis techniques. These include calibration, filtering, subsetting, agglomeration, multi-table comparisons, diversity analysis, parallelized Fast UniFrac, ordination methods, and production of publication-quality graphics; all in a manner that is easy to document, share, and modify. We show how to apply functions from other R packages to phyloseq-represented data, illustrating the availability of a large number of open source analysis techniques. We discuss the use of phyloseq with tools for reproducible research, a practice common in other fields but still rare in the analysis of highly parallel microbiome census data. We have made available all of the materials necessary to completely reproduce the analysis and figures included in this article, an example of best practices for reproducible research. CONCLUSIONS: The phyloseq project for R is a new open-source software package, freely available on the web from both GitHub and Bioconductor.

BACKGROUND: We recently described FastTree, a tool for inferring phylogenies for alignments with up to hundreds of thousands of sequences. Here, we describe improvements to FastTree that improve its accuracy without sacrificing scalability. METHODOLOGY/PRINCIPAL FINDINGS: Where FastTree 1 used nearest-neighbor interchanges (NNIs) and the minimum-evolution criterion to improve the tree, FastTree 2 adds minimum-evolution subtree-pruning-regrafting (SPRs) and maximum-likelihood NNIs. FastTree 2 uses heuristics to restrict the search for better trees and estimates a rate of evolution for each site (the "CAT" approximation). Nevertheless, for both simulated and genuine alignments, FastTree 2 is slightly more accurate than a standard implementation of maximum-likelihood NNIs (PhyML 3 with default settings). Although FastTree 2 is not quite as accurate as methods that use maximum-likelihood SPRs, most of the splits that disagree are poorly supported, and for large alignments, FastTree 2 is 100-1,000 times faster. FastTree 2 inferred a topology and likelihood-based local support values for 237,882 distinct 16S ribosomal RNAs on a desktop computer in 22 hours and 5.8 gigabytes of memory. CONCLUSIONS/SIGNIFICANCE: FastTree 2 allows the inference of maximum-likelihood phylogenies for huge alignments. FastTree 2 is freely available at http://www.microbesonline.org/fasttree.

Advances in modern sequencing technologies allow us to generate sufficient data to analyze hundreds of bacterial genomes from a single machine in a single day. This potential for sequencing massive numbers of genomes calls for fully automated methods to produce high-quality assemblies and variant calls. We introduce Pilon, a fully automated, all-in-one tool for correcting draft assemblies and calling sequence variants of multiple sizes, including very large insertions and deletions. Pilon works with many types of sequence data, but is particularly strong when supplied with paired end data from two Illumina libraries with small e.g., 180 bp and large e.g., 3-5 Kb inserts. Pilon significantly improves draft genome assemblies by correcting bases, fixing mis-assemblies and filling gaps. For both haploid and diploid genomes, Pilon produces more contiguous genomes with fewer errors, enabling identification of more biologically relevant genes. Furthermore, Pilon identifies small variants with high accuracy as compared to state-of-the-art tools and is unique in its ability to accurately identify large sequence variants including duplications and resolve large insertions. Pilon is being used to improve the assemblies of thousands of new genomes and to identify variants from thousands of clinically relevant bacterial strains. Pilon is freely available as open source software.

Advances in next generation technologies have driven the costs of DNA sequencing down to the point that genotyping-by-sequencing (GBS) is now feasible for high diversity, large genome species. Here, we report a procedure for constructing GBS libraries based on reducing genome complexity with restriction enzymes (REs). This approach is simple, quick, extremely specific, highly reproducible, and may reach important regions of the genome that are inaccessible to sequence capture approaches. By using methylation-sensitive REs, repetitive regions of genomes can be avoided and lower copy regions targeted with two to three fold higher efficiency. This tremendously simplifies computationally challenging alignment problems in species with high levels of genetic diversity. The GBS procedure is demonstrated with maize (IBM) and barley (Oregon Wolfe Barley) recombinant inbred populations where roughly 200,000 and 25,000 sequence tags were mapped, respectively. An advantage in species like barley that lack a complete genome sequence is that a reference map need only be developed around the restriction sites, and this can be done in the process of sample genotyping. In such cases, the consensus of the read clusters across the sequence tagged sites becomes the reference. Alternatively, for kinship analyses in the absence of a reference genome, the sequence tags can simply be treated as dominant markers. Future application of GBS to breeding, conservation, and global species and population surveys may allow plant breeders to conduct genomic selection on a novel germplasm or species without first having to develop any prior molecular tools, or conservation biologists to determine population structure without prior knowledge of the genome or diversity in the species.

Outcomes of high-throughput biological experiments are typically interpreted by statistical testing for enriched gene functional categories defined by the Gene Ontology (GO). The resulting lists of GO terms may be large and highly redundant, and thus difficult to interpret.REVIGO is a Web server that summarizes long, unintelligible lists of GO terms by finding a representative subset of the terms using a simple clustering algorithm that relies on semantic similarity measures. Furthermore, REVIGO visualizes this non-redundant GO term set in multiple ways to assist in interpretation: multidimensional scaling and graph-based visualizations accurately render the subdivisions and the semantic relationships in the data, while treemaps and tag clouds are also offered as alternative views. REVIGO is freely available at http://revigo.irb.hr/.

Binary classifiers are routinely evaluated with performance measures such as sensitivity and specificity, and performance is frequently illustrated with Receiver Operating Characteristics (ROC) plots. Alternative measures such as positive predictive value (PPV) and the associated Precision/Recall (PRC) plots are used less frequently. Many bioinformatics studies develop and evaluate classifiers that are to be applied to strongly imbalanced datasets in which the number of negatives outweighs the number of positives significantly. While ROC plots are visually appealing and provide an overview of a classifier's performance across a wide range of specificities, one can ask whether ROC plots could be misleading when applied in imbalanced classification scenarios. We show here that the visual interpretability of ROC plots in the context of imbalanced datasets can be deceptive with respect to conclusions about the reliability of classification performance, owing to an intuitive but wrong interpretation of specificity. PRC plots, on the other hand, can provide the viewer with an accurate prediction of future classification performance due to the fact that they evaluate the fraction of true positives among positive predictions. Our findings have potential implications for the interpretation of a large number of studies that use ROC plots on imbalanced datasets.

As part of a World Health Organization-led effort to update the empirical evidence base for the leishmaniases, national experts provided leishmaniasis case data for the last 5 years and information regarding treatment and control in their respective countries and a comprehensive literature review was conducted covering publications on leishmaniasis in 98 countries and three territories (see 'Leishmaniasis Country Profiles Text S1, S2, S3, S4, S5, S6, S7, S8, S9, S10, S11, S12, S13, S14, S15, S16, S17, S18, S19, S20, S21, S22, S23, S24, S25, S26, S27, S28, S29, S30, S31, S32, S33, S34, S35, S36, S37, S38, S39, S40, S41, S42, S43, S44, S45, S46, S47, S48, S49, S50, S51, S52, S53, S54, S55, S56, S57, S58, S59, S60, S61, S62, S63, S64, S65, S66, S67, S68, S69, S70, S71, S72, S73, S74, S75, S76, S77, S78, S79, S80, S81, S82, S83, S84, S85, S86, S87, S88, S89, S90, S91, S92, S93, S94, S95, S96, S97, S98, S99, S100, S101'). Additional information was collated during meetings conducted at WHO regional level between 2007 and 2011. Two questionnaires regarding epidemiology and drug access were completed by experts and national program managers. Visceral and cutaneous leishmaniasis incidence ranges were estimated by country and epidemiological region based on reported incidence, underreporting rates if available, and the judgment of national and international experts. Based on these estimates, approximately 0.2 to 0.4 cases and 0.7 to 1.2 million VL and CL cases, respectively, occur each year. More than 90% of global VL cases occur in six countries: India, Bangladesh, Sudan, South Sudan, Ethiopia and Brazil. Cutaneous leishmaniasis is more widely distributed, with about one-third of cases occurring in each of three epidemiological regions, the Americas, the Mediterranean basin, and western Asia from the Middle East to Central Asia. The ten countries with the highest estimated case counts, Afghanistan, Algeria, Colombia, Brazil, Iran, Syria, Ethiopia, North Sudan, Costa Rica and Peru, together account for 70 to 75% of global estimated CL incidence. Mortality data were extremely sparse and generally represent hospital-based deaths only. Using an overall case-fatality rate of 10%, we reach a tentative estimate of 20,000 to 40,000 leishmaniasis deaths per year. Although the information is very poor in a number of countries, this is the first in-depth exercise to better estimate the real impact of leishmaniasis. These data should help to define control strategies and reinforce leishmaniasis advocacy.

This paper describes the technical development and accuracy assessment of the most recent and improved version of the SoilGrids system at 250m resolution (June 2016 update). SoilGrids provides global predictions for standard numeric soil properties (organic carbon, bulk density, Cation Exchange Capacity (CEC), pH, soil texture fractions and coarse fragments) at seven standard depths (0, 5, 15, 30, 60, 100 and 200 cm), in addition to predictions of depth to bedrock and distribution of soil classes based on the World Reference Base (WRB) and USDA classification systems (ca. 280 raster layers in total). Predictions were based on ca. 150,000 soil profiles used for training and a stack of 158 remote sensing-based soil covariates (primarily derived from MODIS land products, SRTM DEM derivatives, climatic images and global landform and lithology maps), which were used to fit an ensemble of machine learning methods-random forest and gradient boosting and/or multinomial logistic regression-as implemented in the R packages ranger, xgboost, nnet and caret. The results of 10-fold cross-validation show that the ensemble models explain between 56% (coarse fragments) and 83% (pH) of variation with an overall average of 61%. Improvements in the relative accuracy considering the amount of variation explained, in comparison to the previous version of SoilGrids at 1 km spatial resolution, range from 60 to 230%. Improvements can be attributed to: (1) the use of machine learning instead of linear regression, (2) to considerable investments in preparing finer resolution covariate layers and (3) to insertion of additional soil profiles. Further development of SoilGrids could include refinement of methods to incorporate input uncertainties and derivation of posterior probability distributions (per pixel), and further automation of spatial modeling so that soil maps can be generated for potentially hundreds of soil variables. Another area of future research is the development of methods for multiscale merging of SoilGrids predictions with local and/or national gridded soil products (e.g. up to 50 m spatial resolution) so that increasingly more accurate, complete and consistent global soil information can be produced. SoilGrids are available under the Open Data Base License.

Plastic pollution is ubiquitous throughout the marine environment, yet estimates of the global abundance and weight of floating plastics have lacked data, particularly from the Southern Hemisphere and remote regions. Here we report an estimate of the total number of plastic particles and their weight floating in the world's oceans from 24 expeditions (2007-2013) across all five sub-tropical gyres, costal Australia, Bay of Bengal and the Mediterranean Sea conducting surface net tows (N = 680) and visual survey transects of large plastic debris (N = 891). Using an oceanographic model of floating debris dispersal calibrated by our data, and correcting for wind-driven vertical mixing, we estimate a minimum of 5.25 trillion particles weighing 268,940 tons. When comparing between four size classes, two microplastic <4.75 mm and meso- and macroplastic >4.75 mm, a tremendous loss of microplastics is observed from the sea surface compared to expected rates of fragmentation, suggesting there are mechanisms at play that remove <4.75 mm plastic particles from the ocean surface.

Understanding and interpreting classification decisions of automated image classification systems is of high value in many applications, as it allows to verify the reasoning of the system and provides additional information to the human expert. Although machine learning methods are solving very successfully a plethora of tasks, they have in most cases the disadvantage of acting as a black box, not providing any information about what made them arrive at a particular decision. This work proposes a general solution to the problem of understanding classification decisions by pixel-wise decomposition of nonlinear classifiers. We introduce a methodology that allows to visualize the contributions of single pixels to predictions for kernel-based classifiers over Bag of Words features and for multilayered neural networks. These pixel contributions can be visualized as heatmaps and are provided to a human expert who can intuitively not only verify the validity of the classification decision, but also focus further analysis on regions of potential interest. We evaluate our method for classifiers trained on PASCAL VOC 2009 images, synthetic image data containing geometric shapes, the MNIST handwritten digits data set and for the pre-trained ImageNet model available as part of the Caffe open source package.

Recombination is an important process in the evolution of RNA viruses. For viruses with segmented genomes, recombination can occur by reassortment of entire genomic segments. This process has played a significant role in the adaptation and spread of influenza viruses with novel genomic combinations, i.e. , antigenic shifts. Reassortment is usually identified by discrepancies between the topologies of trees reconstructed from different genomic segments, e.g. , subtree prune and regraft (SPR) rearrangements. However, this phylogenetic discordance can also result from errors in tree reconstruction due to insufficient information. The objective of this study was to evaluate the contribution of phylogenetic reconstruction error to the inference of reassortment events from influenza A virus (IAV) genomes from avian hosts. We curated a dataset of 11,795 complete avian H5Nx IAV genomes and sampled subsets of HA sequences with different expected levels of reconstruction error. Next, we simulated molecular evolution along the trees reconstructed from these HA samples, varying rates of evolution and sequence lengths to correspond to the other genomic segments. Because these simulations had no reassortment, any phylogenetic discordance identified from the resulting trees must be caused by reconstruction error. We found that simulations from random samples yielded about 32% as many SPRs as obtained from real data. This false positive rate was reduced to 4% when simulating trees from the most genetically divergent subset of HA sequences. By modifying the simulated trees with SPR rearrangements, we observed a false negative rate of 8.5%, which increased with the number of SPRs per tree. Overall, we found that reassortment can be substantially overestimated from phylogenetic discordance, underscoring the need to account for phylogenetic uncertainty and sampling design when analyzing IAV datasets for reassortment.

Deep-learning-based holographic reconstruction requires numerical data with reliable ground truth for training and quantitative evaluation. Experimental hologram acquisition and pixel-level phase annotation can be difficult to control, making physically defined numerical cases useful as reproducible benchmarks. Here, three analytically specified phase targets—a smooth Gaussian distribution, a four-level step distribution, and a radially structured Siemens Star—are used to generate off-axis digital holograms with known phase ground truth. A frequency-domain filtering procedure is applied before inverse Fourier reconstruction, followed by DCT-based global phase unwrapping. Reconstruction quality is quantified using root mean square error (RMSE), mean absolute error (MAE), correlation coefficient, and profile-based error analysis. The results show that reconstruction errors are comparatively small in smooth regions and increase near abrupt or high-spatial-frequency phase transitions. The three cases are therefore presented as controlled benchmark examples for algorithm verification and baseline comparison, not as a complete substitute for experimental holographic datasets or as a complete training corpus for deep-learning models.

Economic diversification is commonly modeled using co-occurrence-based measures of relatedness that quantify pairwise overlap among activities in a capability network. Although such measures are derived from network structure, they operate at the level of first-order neighborhood relations and summarize proximity through local co-occurrence. They do not explicitly encode higher-order structural context, such as indirect connectivity patterns or global positional similarity, into a continuous representation of entities and activities. We develop an embedding-based framework that learns such representations from network topology, enabling the construction of interpretable geometric features, including capability alignment, displacement, and portfolio entropy, that extend standard measures used in economic complexity. We apply this framework to two systems: international trade (1998–2023) and public procurement firms (2009–2024). The resulting representations reveal different structural organization across scales: firms exhibit modular clustering patterns, whereas countries occupy broader, nested capability configurations. In predictive tasks for product and activity specialization, embedding-derived features provide additional explanatory power beyond relatedness in several settings, particularly when forecasting variation in the magnitude of comparative advantage. For country-level entry events, however, persistence remains the strongest predictor. Our results show that relatedness remains the dominant predictor of entry events. In contrast, embedding-derived features become more informative in specific settings, particularly for modeling variation in specialization magnitude and in low-relatedness regimes. More broadly, the findings indicate that diversification depends on structural signals that are both prediction-target-dependent and scale-dependent, with relatedness and embedding-derived features capturing different aspects of the capability space (local overlap vs. structural position).

Background In sub-Saharan Africa, child mortality remains a critical public health challenge, with Zambia exhibiting some of the highest rates in the region. While survival analysis has identified key risk factors, these methods often fail to capture complex non-linear interactions and the dynamic, time-varying nature of longitudinal markers like birth intervals. This study aimed to develop and compare predictive survival models to improve risk stratification for under-five mortality using the 2024 Zambia Demographic and Health Survey (ZDHS). Methods A cohort of 35,610 live births was initially identified, with a final analytic sample of 9,362 singleton births and 1,236 mortality events (13.2%) after applying exclusion criteria for maternal age and data completeness. We compared four modeling frameworks for predictive comparison: Cox Proportional Hazards (Cox PH), Lasso Cox regression, Random Survival Forests (RSF), and Gradient Boosting Machines (GBM), alongside a Bayesian Joint Model for inference on the birth interval and mortality association. The proportional hazards assumption was assessed using Schoenfeld residuals, and stratified Cox models were employed to address time-varying effects. Model performance was evaluated using nested cross-validation with mother-level outer folds and time-dependent Concordance Indices (C-index) and Area Under the Curve (AUC) metrics across a 48-month follow-up period. Results Under nested cross-validation, the four survival models demonstrated comparable predictive performance, with C-indices ranging from 0.745 (Cox PH) to 0.750 (Gradient Boosting Machine) and Integrated Brier Scores ranging from 0.095 to 0.096; pairwise comparisons across models were not statistically significant (DeLong’s test: all p > 0.05). The Bayesian Joint Model revealed a significant association between the longitudinal birth interval process and the hazard of death ( α = 0.035 , 95% CI: 0.023–0.048, p < 0.001), supporting the maternal depletion hypothesis. Variable importance analysis identified birth interval and maternal age as the most influential predictors of child survival, followed by previous child deaths. Stratified analysis of the Cox model revealed significant effects for maternal education (Secondary: HR = 0.426, 95% CI: 0.318–0.570) and previous child deaths (HR = 1.648, 95% CI: 1.529–1.777). Conclusion Advanced survival modeling using joint modeling techniques provides deeper biological insights into child mortality dynamics, and machine learning and traditional survival models offered equivalent predictive discrimination and calibration in this analysis. These findings support targeted public health interventions focusing on birth spacing, maternal age, and enhanced monitoring for high-risk families as critical strategies for reducing under-five mortality in Zambia.

Chronic obstructive pulmonary disease (COPD) is a complex and common condition with significant health, social, and economic burdens, especially due to frequent hospitalizations and comorbidities. International research shows current care for COPD is limited by delayed diagnosis, poor coordination among providers, and inconsistent adherence to treatment guidelines, which negatively impacts patient outcomes. Clinical pathways offer a promising solution by improving care coordination, aligning treatment with evidence-based guidelines, and enhancing both patient experiences and system efficiency; however, questions remain about how, for whom, and in which contexts they are most effective. Using a realist evaluation approach, our study will evaluate how, for whom, and in which circumstances an integrated clinical pathway for COPD works in Saskatchewan, Canada. First, we will develop initial program theories based on literature and stakeholder input. Second, we will test these theories using administrative data and realist interviews with healthcare staff, managers, providers, COPD patients, and caregivers. Third, we will refine our theories based on data collected. Fourth, we will translate our findings to enhance future implementation. These phases will be repeated over a three-year span to continuously improve pathway implementation processes and monitor improvements in patient outcomes. Findings will be shared through integrated knowledge translation with system decision-makers and patients. Findings will also be disseminated to the broader research community through conference presentations and peer-reviewed publications.

The main purpose of this study was to determine the differences in executive function performance levels of esports players who play First-Person Shooter (FPS) and Multiplayer Online Battle Arena (MOBA) game types compared to non-video game players (NVGPs). Other purposes of the study are to examine the relationships of executive function components and their relationships with in-game rank and total game play time variables. The study groups consisted of ninety volunteer participants, all of whom were male young adults between the ages of 18 and 34, 30 of whom actively played FPS video games, 30 of whom actively played MOBA video games, and 30 of whom NVGPs. Corsi Block-Tapping Test (CBT), Wisconsin (Berg) Card Sorting Test (WCST), and Stroop Test Çapa Form (ST) were used as data collection tools. According to the findings of the study; FPS players had significantly higher scores than NVGPs (p < 0.001, F = 8.55) in CBT and less error percentage than the NVGPs (p < 0.01, F = 5.58) and MOBA players had significantly less interference time than NVGPs in the ST (p < 0.001, F = 7.66). In addition, there was a significant negative correlation between CBT and ST interference times for all players (p < 0.05; r = −0.28) and a significant negative correlation between in-game skill and ST interference times for FPS players (p < 0.05; r = −0.41). These findings suggest that digital video game experience shapes cognitive profiles based on genre-specific demands; while FPS gameplay primarily enhances visuospatial working memory capacity, MOBA gameplay fosters stronger inhibition control, with higher in-game skill level reflecting superior inhibitory control in FPS gameplay, and relationship of visuospatial working memory and inhibition components demonstrating an underlying interrelated structure.

Background White feces syndrome (WFS) is a major health problem affecting intensive shrimp aquaculture, but its underlying mechanisms remain unclear. Increasing evidence suggests that WFS may involve complex interactions among host physiological status, intestinal microbial dysbiosis, and environmental microbial communities. However, the relationships among these factors within commercial shrimp culture systems remain poorly understood. Objectives This study aimed to investigate the associations among intestinal antioxidant capacity, intestinal microbiota, sediment microbial communities, and water quality parameters in Litopenaeus vannamei affected by WFS under small-pond culture conditions. Methods Healthy and WFS-affected shrimp were collected from a commercial shrimp farm in Zhanjiang, Guangdong Province, China. Intestinal antioxidant parameters, including total antioxidant capacity (T-AOC), superoxide dismutase (SOD), glutathione peroxidase (GSH-Px), and malondialdehyde (MDA), were determined. Intestinal and sediment microbial communities were analyzed using 16S rRNA gene sequencing. Water quality parameters and microbial community composition were compared between healthy and WFS groups. Results Compared with healthy shrimp, WFS-affected shrimp exhibited significantly increased intestinal MDA levels and decreased activities of T-AOC, SOD, and GSH-Px, indicating an imbalance in antioxidant status. Significant differences in intestinal microbial community composition were observed between groups, as confirmed by beta-diversity analysis (PERMANOVA, R 2 = 0.407, P = 0.022). WFS-associated microbial alterations included changes in dominant bacterial taxa and predicted microbial functional potential related to metabolism. The dominant taxa, referring to the most abundant microbial groups based on relative abundance, differed between healthy and WFS-affected shrimp. Sediment microbial communities also showed substantial compositional shifts, suggesting alterations in the surrounding microbial environment. In addition, nitrate and total nitrogen concentrations were significantly reduced in the WFS group. Conclusion These findings indicate that WFS occurrence is associated with coordinated alterations in shrimp antioxidant status, intestinal microbial structure, sediment microbial communities, and environmental conditions. This study provides new insights into WFS from a host–microbiota–environment interaction perspective and highlights potential indicators for monitoring shrimp health in aquaculture systems.

Introduction This study aimed to assess global stakeholder preferences for key attributes of point-of-care (POC) HIV Nucleic Acid Amplification Tests (NAAT), with the aim of informing product development to support progress toward UNAIDS 95-95-95 targets. Methods A web-based conjoint survey was conducted in August through November 2024 among HIV clinicians, laboratory professionals, and researchers. Participants were asked to evaluate hypothetical POC HIV NAAT profiles varying by six attributes: sensitivity / limit of detection (LOD), sample-collection method, turnaround time (TAT), result format, price, and power source. Relative importance and part-worth utilities were calculated to assess trade-offs in decision-making. Results One hundred three respondents from 33 countries, primarily in Africa (75%), completed the survey. The relative importance placed on the attributes in the conjoint task varied based on the setting. In community and home-based settings, participants prioritized fingerstick sample collection (25.90%) and faster TATs (28.8%), even at the expense of sensitivity (15.70%). By contrast, respondents from hospitals and health centers valued tests with high sensitivity (42.6%, 40.9%) and faster TATs (31.5%, 22.6%). Across all respondents, price (7.0%), result format (3.30%), and power source (7.25%) were less influential. Conclusions Stakeholders were willing to make trade-offs among desirable attributes, prioritizing rapid turnaround time, strong sensitivity, and fingerstick sampling over affordability, power requirements, and result format in the conjoint task. These preferences vary by level of care and by use case, underscoring the importance of aligning product features with contextual needs. Engaging stakeholders early in the design process is critical to ensure new POC HIV NAAT technologies are both acceptable and scalable in low-resource settings.

Infrared object detection plays an important role in low-light and complex-environment perception. However, infrared images often exhibit low signal-to-noise ratio, weak target-background contrast, and limited texture information, which can degrade detection performance in small-target and cluttered-background scenarios while making it difficult to balance detection accuracy and model complexity. To address these challenges, this paper proposes TCLNet, a Thermal Contrast-Aware Lightweight Network for infrared object detection. TCLNet introduces a Thermal Contrast Encoder (TCE) to enhance local contrast-sensitive target representations and a Thermal-Guided Cross-Feature Attention (TGCA) mechanism to promote interactions between contrast-enhanced and semantic features, thereby improving target discrimination and suppressing background interference.Experiments were conducted independently on the M3FD, FLIR ADAS v2 (denoted as FLIR_2), and NUDT-SIRST1 datasets. TCLNet achieves mAP50 values of 60.5%, 45.4%, and 91.4%, respectively, corresponding to improvements of 12.8, 11.3, and 7.5 percentage points over the YOLO12n baseline. Meanwhile, the model maintains a compact size of approximately 2M parameters. These results demonstrate that TCLNet improves infrared target detection, particularly in low-SNR and small-target conditions, while achieving a favorable balance between detection accuracy and model complexity.