Search

Abstract Rectal cancer ranks as the second leading cause of cancer-related deaths. Neoadjuvant therapy for rectal cancer patients often results in individuals that respond well to therapy and those that respond poorly, requiring life-altering excision surgery. It is inadequately understood what dictates this responder/nonresponder divide. Our major aim is to identify what factors in the tumor microenvironment drive a fraction of rectal cancer patients to respond to radiotherapy. We also sought to distinguish potential biomarkers that would indicate a positive response to therapy and design combinatorial therapeutics to enhance radiotherapy efficacy. To address this, we developed an orthotopic murine model of rectal cancer treated with short course radiotherapy that recapitulates the bimodal response observed in the clinic. We utilized a robust combination of transcriptomics and protein analysis to identify differences between responding and nonresponding tumors. Our mouse model recapitulates human disease in which a fraction of tumors respond to radiotherapy (responders) while the majority are nonresponsive. We determined that responding tumors had increased damage-induced cell death, and a unique immune-activation signature associated with tumor-associated macrophages, cancer-associated fibroblasts, and CD8+ T cells. This signature was dependent on radiation-induced increases of Type I Interferons (IFNs). We investigated a therapeutic approach targeting the cGAS/STING pathway and demonstrated improved response rate following radiotherapy. These results suggest that modulating the Type I IFN pathway has the potential to improve radiation therapy efficacy in RC.

ABSTRACT HIV incidence is a key measure for tracking disease spread and identifying populations and geographic regions where new infections are most concentrated. The HIV sequence population provides a robust signal for the stage of infection. Large-scale and high-precision HIV sequencing is crucial for effective genomic incidence surveillance. We produced 1,034 full-length envelope gene sequences from a seroconversion cohort by conducting HIV microdrop sequencing and measuring the genomic incidence assay’s genome similarity index (GSI) dynamics. The measured dynamics of 9 of 12 individuals aligned with the GSI distribution estimated independently using 417 publicly available incident samples. We enhanced the capacity to identify individuals with recent infections, achieving predicted detection accuracies of 92% (89%–94%) for cases within 6 months and 81% (74%–87%) for cases within 9 months. These accuracy levels agreed with the observed detection accuracy intervals of an independent validation data set. Additionally, we produced 131 full-length envelope gene sequences from eight individuals with chronic HIV infection. This analysis confirmed a false recency rate (FRR) of 0%, which was consistent with 162 publicly available chronic samples. The mean duration of recent infection (MDRI) was 238 (209–267) days, indicating an 83% improvement in performance compared to current recent infection testing algorithms. The shifted Poisson mixture model was then used to estimate the time since infection, and the model estimates showed an 88% consistency with the days post infection derived from HIV RNA test dates and/or seroconversion dates. HIV microdrop sequencing provides unique prospects for large-scale incidence surveillance using high-throughput sequencing. Importance Accurate identification of recently infected individuals is vital for prioritizing specific populations for interventions, reducing onward transmission risks, and optimizing public health services. However, current HIV-specific antibody-based methods have not been satisfactory in accurately identifying incident cases, hindering the use of HIV recency testing for prevention efforts and partner protection. Genomic incidence assays offer a promising alternative for identifying recent infections. In our study, we used microdroplet technologies to produce a large number of complete HIV envelope gene sequences, enabling the accurate detection of early infection signs. We assessed the dynamics of the incidence assay’s metrics and compared them with statistical models. Our approach demonstrated high accuracy in identifying individuals with recent infections, achieving predicted detection rates exceeding 90% within 6 months and over 80% within 9 months of infection. This high-resolution method holds significant potential for enhancing the effectiveness of HIV incidence screening for case-based surveillance in public health initiatives.

Abstract Background Quantitative real-time PCR (qPCR) is one of the most widely used methods to measure gene expression. An important aspect of qPCR data that has been largely ignored is the presence of non-detects: reactions failing to exceed the quantification threshold and therefore lacking a measurement of expression. While most current software replaces these non-detects with a value representing the limit of detection, this introduces substantial bias in the estimation of both absolute and differential expression. Single imputation procedures, while an improvement on previously used methods, underestimate residual variance, which can lead to anti-conservative inference. Results We propose to treat non-detects as non-random missing data, model the missing data mechanism, and use this model to impute missing values or obtain direct estimates of model parameters. To account for the uncertainty inherent in the imputation, we propose a multiple imputation procedure, which provides a set of plausible values for each non-detect. We assess the proposed methods via simulation studies and demonstrate the applicability of these methods to three experimental data sets. We compare our methods to mean imputation, single imputation, and a penalized EM algorithm incorporating non-random missingness (PEMM). The developed methods are implemented in the R/Bioconductor package nondetects. Conclusions The statistical methods introduced here reduce discrepancies in gene expression values derived from qPCR experiments in the presence of non-detects, providing increased confidence in downstream analyses.

Fecal contamination of surface water has been associated with multiple enteric disease outbreaks and food recalls. Thus, it is important to understand factors associated with fecal contamination of agricultural water sources. Since fecal indicator bacteria (FIB) were used to monitor surface water for potential fecal contamination, the purpose of the present study was to characterize associations between environmental factors, and (i) FIB (E. coli, Enterococcus, and coliform) levels, and (ii) host-specific fecal marker detection. This study used data collected from 224 sites along 3 waterways, which spanned an urban-rural gradient around Syracuse, New York. Between 2008 and 2017, 2,816 water samples were collected, and E. coli, Enterococcus, and/or coliform concentrations were enumerated. Thirty-one samples were also tested for human and ruminant microbial source-tracking markers. Water quality (e.g., turbidity, nitrate) and weather data were also collected for each site. Univariable Bayesian regression was used to characterize the relationship between each microbial target and land use, water quality, and weather factor. For each model, probability of direction and region of practical equivalence overlap (ROPE) were calculated to characterize the association's direction and strength, respectively. While levels of different FIB were not correlated with each other, FIB levels were associated with environmental conditions. Specifically, FIB levels were also positively associated with temperature, nutrient and sediment levels. Log10 E. coli levels increased by 0.20 (CI = 0.11, 0.31) and log10 Enterococcus levels increased by 0.68 (CI = 0.08, 1.24) for each log10 increase in salinity and nitrate, respectively. These findings may indicate that similar processes drove microbial, sediment, and nutrient contamination of the sampled watersheds. While fecal contamination was strongly associated with land use, the direction of association varied between FIBs and the buffer distance used to calculate land use metrics. E. coli levels and human marker detection were positively associated with percent pasture cover within 122, 366, and 1,098 m of the sampling site, while Enterococcus and coliform levels were only associated with pasture cover within 1,098 m (not 122 or 366 m). Ruminant markers were positively associated with pasture cover within 122 m, but not 366 or 1,098 m. These findings highlight the importance of considering (i) adjacent land use (and associated non-point sources of contamination) when developing strategies for managing fecal hazards associated in agricultural and recreational water, and (ii) spatial scale (e.g., 122 vs. 1,098 m) when developing these strategies.

Recent studies have shown that predictive models can supplement or provide alternatives to E. coli-testing for assessing the potential presence of food safety hazards in water used for produce production. However, these studies used balanced training data and focused on enteric pathogens. As such, research is needed to determine 1) if predictive models can be used to assess Listeria contamination of agricultural water, and 2) how resampling (to deal with imbalanced data) affects performance of these models. To address these knowledge gaps, this study developed models that predict nonpathogenic Listeria spp. (excluding L. monocytogenes) and L. monocytogenes presence in agricultural water using various combinations of learner (e.g., random forest, regression), feature type, and resampling method (none, oversampling, SMOTE). Four feature types were used in model training: microbial, physicochemical, spatial, and weather. “Full models” were trained using all four feature types, while “nested models” used between one and three types. In total, 45 full (15 learners*3 resampling approaches) and 108 nested (5 learners*9 feature sets*3 resampling approaches) models were trained per outcome. Model performance was compared against baseline models where E. coli concentration was the sole predictor. Overall, the machine learning models outperformed the baseline E. coli models, with random forests outperforming models built using other learners (e.g., rule-based learners). Resampling produced more accurate models than not resampling, with SMOTE models outperforming, on average, oversampling models. Regardless of resampling method, spatial and physicochemical water quality features drove accurate predictions for the nonpathogenic Listeria spp. and L. monocytogenes models, respectively. Overall, these findings 1) illustrate the need for alternatives to existing E. coli-based monitoring programs for assessing agricultural water for the presence of potential food safety hazards, and 2) suggest that predictive models may be one such alternative. Moreover, these findings provide a conceptual framework for how such models can be developed in the future with the ultimate aim of developing models that can be integrated into on-farm risk management programs. For example, future studies should consider using random forest learners, SMOTE resampling, and spatial features to develop models to predict the presence of foodborne pathogens, such as L. monocytogenes, in agricultural water when the training data is imbalanced.

Since E. coli is considered a fecal indicator in surface water, government water quality standards and industry guidance often rely on E. coli monitoring to identify when there is an increased risk of pathogen contamination of water used for produce production (e.g., for irrigation). However, studies have indicated that E. coli testing can present an economic burden to growers and that time lags between sampling and obtaining results may reduce the utility of these data. Models that predict E. coli levels in agricultural water may provide a mechanism for overcoming these obstacles. Thus, this proof-of-concept study uses previously published datasets to train, test, and compare E. coli predictive models using multiple algorithms and performance measures. Since the collection of different feature data carries specific costs for growers, predictive performance was compared for models built using different feature types [geospatial, water quality, stream traits, and/or weather features]. Model performance was assessed against baseline regression models. Model performance varied considerably with root-mean-squared errors and Kendall’s Tau ranging between 0.37 and 1.03, and 0.07 and 0.55, respectively. Overall, models that included turbidity, rain, and temperature outperformed all other models regardless of the algorithm used. Turbidity and weather factors were also found to drive model accuracy even when other feature types were included in the model. These findings confirm previous conclusions that machine learning models may be useful for predicting when, where, and at what level E. coli (and associated hazards) are likely to be present in preharvest agricultural water sources. This study also identifies specific algorithm-predictor combinations that should be the foci of future efforts to develop deployable models (i.e., models that can be used to guide on-farm decision-making and risk mitigation). When deploying E. coli predictive models in the field, it is important to note that past research indicates an inconsistent relationship between E. coli levels and foodborne pathogen presence. Thus, models that predict E. coli levels in agricultural water may be useful for assessing fecal contamination status and ensuring compliance with regulations but should not be used to assess the risk that specific pathogens of concern (e.g., Salmonella, Listeria) are present.

While the Food Safety Modernization Act established standards for the use of surface water for produce production, water quality is known to vary over space and time. Targeted approaches for identifying hazards in water that account for this variation may improve growers' ability to address pre-harvest food safety risks. Models that utilize publicly-available data (e.g., land-use, real-time weather) may be useful for developing these approaches. The objective of this study was to use pre-existing datasets collected in 2017 (N = 181 samples) and 2018 (N = 191 samples) to train and test models that predict the likelihood of detecting Salmonella and pathogenic E. coli markers (eaeA, stx) in agricultural water. Four types of features were used to train the models: microbial, physicochemical, spatial and weather. “Full models” were built using all four features types, while “nested models” were built using between one and three types. Twenty learners were used to develop separate full models for each pathogen. Separately, to assess information gain associated with using different feature types, six learners were randomly selected and used to develop nine, nested models each. Performance measures for each model were then calculated and compared against baseline models where E. coli concentration was the sole covariate. In the methods, we outline the advantages and disadvantages of each learner. Overall, full models built using ensemble (e.g., Node Harvest) and “black-box” (e.g., SVMs) learners out-performed full models built using more interpretable learners (e.g., tree- and rule-based learners) for both outcomes. However, nested eaeA-stx models built using interpretable learners and microbial data performed almost as well as these full models. While none of the nested Salmonella models performed as well as the full models, nested models built using spatial data consistently out-performed models that excluded spatial data. These findings demonstrate that machine learning approaches can be used to predict when and where pathogens are likely to be present in agricultural water. This study serves as a proof-of-concept that can be built upon once larger datasets become available and provides guidance on the learner-data combinations that should be the foci of future efforts (e.g., tree-based microbial models for pathogenic E. coli).

Abstract HIV incidence is a primary metric for epidemic surveillance and prevention efficacy assessment. HIV incidence assay performance is evaluated via false recency rate (FRR) and mean duration of recent infection (MDRI). We conducted a meta-analysis of 438 incident and 305 chronic specimens’ HIV envelope genes from a diverse global cohort. The genome similarity index (GSI) accurately characterized infection stage across diverse host and viral factors. All except one chronic specimen had GSIs below 0.67, yielding a FRR of 0.33 [0-0.98] %. We modeled the incidence assay biomarker dynamics with a logistic link function assuming individual variabilities in a Beta distribution. The GSI probability density function peaked close to 1 in early infection and 0 around two years post infection, yielding MDRI of 420 [361, 467] days. We tested the assay by newly sequencing 744 envelope genes from 59 specimens of 21 subjects who followed from HIV negative status. Both standardized residuals and Anderson-Darling tests showed that the test dataset was statistically consistent with the model biomarker dynamics. This is the first reported incidence assay meeting the optimal FRR and MDRI performance standards. Signatures of HIV gene diversification can allow precise cross-sectional surveillance with a desirable temporal range of incidence detection.