Zheng Xu

Public-health respiratory-disease surveillance requires distinguishing expected seasonal illness activity from atypical changes that may warrant further investigation. Weekly influenza-like illness (ILI) data support situational monitoring, retrospective anomaly screening, and short-term forecasting, but these tasks are often conducted using separate statistical baselines. We empirically evaluated whether seasonal–trend decomposition using LOESS (STL) could provide a common interpretable baseline using weekly national U.S. ILI data from 1997 to 2025. Expected activity was represented by the estimated trend plus recurring seasonal component, while residuals represented baseline-adjusted deviations. Residual exceedance and persistence summaries were examined retrospectively, and forecast performance was evaluated through rolling-origin comparisons with SARIMA, seasonal naive, and ETS benchmarks at 4-, 8-, and 12-week horizons. The decomposition highlighted major departures from the recurring annual pattern, including the 2009 H1N1 pandemic and COVID-19-era disruptions. At the 4-week horizon, STL had the lowest RMSE (0.925), slightly below SARIMA (0.929), seasonal naive (1.270), and ETS (1.710). At the 8- and 12-week horizons, seasonal naive performed best, with RMSEs of 1.270 and 1.280, compared with 1.280 and 1.450 for STL. The two-week persistence rule produced no pre-peak warnings at any tested threshold, indicating that retrospective residual screening did not by itself provide effective early warning. The empirical study demonstrates that STL provides a transparent common baseline for linking monitoring, retrospective anomaly screening, and short-term forecasting while distinguishing expected seasonal progression from baseline-adjusted deviations.

Logistic regression is widely used in complex data analysis. When predictors are at individual level and the response at aggregate level, logistic regression can be estimated using the Maximum Likelihood Estimation (MLE) method with the joint likelihood function formed by Poisson binomial distributions. When directly maximizing the complicated likelihood function, the performance of MLE will worsen as the number of predictors increases. In this article, we propose an expectation-maximization (EM) algorithm to avoid the direct maximization of the complicated likelihood function. Simulation studies have been conducted to evaluate the performance of our EM estimator compared to different estimators proposed in the literature. Two real data-based studies have been conducted to illustrate the use of the different estimators. Our EM estimator proves efficient for the logistic regression problem with an aggregate-level response and individual-level predictors.

This study aimed to understand the changes in the milk and gut microbiota of dairy cows with mastitis, and to further explore the relationship between mastitis and the microbiota. In this study, we extracted microbial DNA from healthy and mastitis cows and performed high-throughput sequencing using the Illumina NovaSeq sequencing platform. OTU clustering was performed to analyze complexity, multi-sample comparisons, differences in community structure between groups, and differential analysis of species composition and abundance. The results showed that there were differences in microbial diversity and community composition in the milk and feces of normal and mastitis cows, where the diversity of microbiota decreased and species abundance increased in the mastitis group. There was a significant difference in the flora composition of the two groups of samples (P < 0.05), especially at the genus level, the difference in the milk samples was Sphingomonas (P < 0.05) and Stenotrophomonas (P < 0.05), the differences in stool samples were Alistipes (P < 0.05), Flavonifractor (P < 0.05), Agathobacter (P < 0.05) and Pygmaiobacter (P < 0.05). In conclusion, the microbiota of the udder and intestinal tissues of dairy cows suffering from mastitis will change significantly. This suggests that the development of mastitis is related to the endogenous pathway of microbial intestinal mammary glands, but the mechanisms involved need further study.

To study the relationship between genetic variants and phenotypes, association testing is adopted; however, most association studies are conducted by genotype-based testing. Testing methods based on next-generation sequencing (NGS) data without genotype calling demonstrate an advantage over testing methods based on genotypes in the scenarios when genotype estimation is not accurate. Our objective was to develop NGS data-based methods for association studies to fill the gap in the literature. Single-variant testing methods based on NGS data have been proposed, including our previously proposed single-variant NGS data-based testing method, i.e., UNC combo method. The NGS data-based group testing method has been proposed by us using a linear model framework which can handle continuous responses. In this paper, we extend our linear model-based framework to a generalized linear model-based framework so that the methods can handle other types of responses especially binary responses which is a common problem in association studies. To evaluate the performance of various estimators and compare them we performed simulation studies. We found that all methods have Type I errors controlled, and our NGS data-based methods have better performance than genotype-based methods for other types of responses, including binary responses (logistics regression) and count responses (Poisson regression), especially when sequencing depth is low. We have extended our previous linear model (LM) framework to a generalized linear model (GLM) framework and derived NGS data-based methods for a group of genetic variables. Compared with our previously proposed LM-based methods, the new GLM-based methods can handle more complex responses (for example, binary responses and count responses) in addition to continuous responses. Our methods have filled the literature gap and shown advantage over their corresponding genotype-based methods in the literature.

Nonprobability samples have been used frequently in practice including public health study, economics, education, and political polls. Naïve estimates based on nonprobability samples without any further adjustments may suffer from serious selection bias. Mass imputation has been shown to be effective in practice to improve the representativeness of nonprobability samples. It builds an imputation model based on nonprobability samples and generates imputed values for all units in the probability samples. In this paper, we compare two mass imputation approaches including latent joint multivariate normal model mass imputation (e.g., Generalized Efficient Regression-Based Imputation with Latent Processes (GERBIL)) and fully conditional specification (FCS) procedures for integrating multiple outcome variables simultaneously. The Monte Carlo simulation study shows the benefits of GERBIL and FCS with predictive mean matching in terms of balancing the Monte Carlo bias and variance. We further evaluate our proposed method by combining the information from Tribal Behavioral Risk Factor Surveillance System and Behavioral Risk Factor Surveillance System data files.

Deep learning has become widely used in image analysis. Transfer learning can make use of information from other datasets for the analysis of the chosen dataset. When there is a small number of images at hand, transfer learning using pre-trained models with coefficients already estimated from other datasets is recommended. This is in contrast to deep learning, where most model parameters are re-estimated. Deep transfer learning uses pre-trained models with fixed weight parameters in the lower layers; thus, deep learning can be viewed as a two-stage approach: (1) feature extraction from a lower neural network layer and (2) estimation of a neural network using the extracted features as inputs. Since deep transfer learning is feature extraction, we can extend the two-stage approach to a more general two-stage framework: (1) feature extraction using multiple methods and (2) machine learning methods taking extracted features as inputs. We evaluate the performance of methods with different Stage 1 and Stage 2 approaches in predicting the phenotype leaf numbers based on a multi-view plant imaging dataset. This paper contains an evaluation of different two-stage machine learning methods for multi-view image data in plant image phenotyping.

Association testing has been widely used to study the relationship between phenotypes and genetic variants. Most testing methods are based on genotypes. To avoid genotype calling and directly test on next-generation sequencing (NGS) data, sequencing data-based methods have been proposed and shown advantages over genotype-based testing methods in scenarios where genotype calling is inaccurate. Most sequencing data-based testing methods are based on a single genetic marker. The objective of this paper is to extend the methods to allow testing for the association of a continuous response variable with a group of common variants or a group of rare variants without genotype calling. Our proposed methods are derived based on a standard linear model framework. We derive the joint significant test (JS) for a group of common genetic variables and the variable collapse test (VC) for a group of rare genetic variables. We have conducted extensive simulation studies to evaluate the performance of different estimators. According to our results, we found (1) all methods, including our proposed NGS data-based methods and genotype-based methods, can control the Type I error rate probability well; (2) our proposed NGS data-based methods can achieve better performance in terms of statistical power compared with their corresponding genotype-based methods in the literature; (3) when sequencing depth increases, the performance of all methods increases, and the difference between the performance of NGS data-based methods and corresponding genotype-based methods decreases. In conclusion, we have proposed NGS data-based methods that allow testing for the significance of a group of variants using a linear model framework and have shown the advantage of our NGS data-based methods over genotype-based methods in the literature.

We propose estimation methods to conduct logistic regression based on individual-level predictors and aggregate-level responses. We derive the likelihood of logistic models in this situation and proposed estimators with different optimization methods. Simulation studies have been conducted to evaluate and compare the performance of the different estimators. A real data-based study has been conducted to illustrate the use of our estimators and compare the different estimators.