Browsing by Subject "Longitudinal data"

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Interep: An R Package for High-Dimensional Interaction Analysis of the Repeated Measurement Data
    (MDPI, 2022-03-19) Zhou, Fei; Ren, Jie; Liu, Yuwen; Li, Xiaoxi; Wang, Weiqun; Wu, Cen; Biostatistics and Health Data Science, School of Medicine
    We introduce interep, an R package for interaction analysis of repeated measurement data with high-dimensional main and interaction effects. In G × E interaction studies, the forms of environmental factors play a critical role in determining how structured sparsity should be imposed in the high-dimensional scenario to identify important effects. Zhou et al. (2019) (PMID: 31816972) proposed a longitudinal penalization method to select main and interaction effects corresponding to the individual and group structure, respectively, which requires a mixture of individual and group level penalties. The R package interep implements generalized estimating equation (GEE)-based penalization methods with this sparsity assumption. Moreover, alternative methods have also been implemented in the package. These alternative methods merely select effects on an individual level and ignore the group-level interaction structure. In this software article, we first introduce the statistical methodology corresponding to the penalized GEE methods implemented in the package. Next, we present the usage of the core and supporting functions, which is followed by a simulation example with R codes and annotations. The R package interep is available at The Comprehensive R Archive Network (CRAN).
  • Thumbnail Image
    Item
    Multi-task deep autoencoder to predict Alzheimer's disease progression using temporal DNA methylation data in peripheral blood
    (Elsevier, 2022-10-23) Chen, Li; Saykin, Andrew J.; Yao, Bing; Zhao, Fengdi; Alzheimer’s Disease Neuroimaging Initiative (ADNI); Radiology and Imaging Sciences, School of Medicine
    Traditional approaches for diagnosing Alzheimer's disease (AD) such as brain imaging and cerebrospinal fluid are invasive and expensive. It is desirable to develop a useful diagnostic tool by exploiting biomarkers obtained from peripheral tissues due to their noninvasive and easily accessible characteristics. However, the capacity of using DNA methylation data in peripheral blood for predicting AD progression is rarely known. It is also challenging to develop an efficient prediction model considering the complex and high-dimensional DNA methylation data in a longitudinal study. Here, we develop two multi-task deep autoencoders, which are based on the convolutional autoencoder and long short-term memory autoencoder to learn the compressed feature representation by jointly minimizing the reconstruction error and maximizing the prediction accuracy. By benchmarking on longitudinal DNA methylation data collected from the peripheral blood in Alzheimer's Disease Neuroimaging Initiative, we demonstrate that the proposed multi-task deep autoencoders outperform state-of-the-art machine learning approaches for both predicting AD progression and reconstructing the temporal DNA methylation profiles. In addition, the proposed multi-task deep autoencoders can predict AD progression accurately using only the historical DNA methylation data and the performance is further improved by including all temporal DNA methylation data. Availability:: https://github.com/lichen-lab/MTAE.
  • Thumbnail Image
    Item
    Sparse group variable selection for gene-environment interactions in the longitudinal study
    (Wiley, 2022) Zhou, Fei; Lu, Xi; Ren, Jie; Fan, Kun; Ma, Shuangge; Wu, Cen; Biostatistics and Health Data Science, School of Medicine
    Penalized variable selection for high dimensional longitudinal data has received much attention as it can account for the correlation among repeated measurements while providing additional and essential information for improved identification and prediction performance. Despite the success, in longitudinal studies, the potential of penalization methods is far from fully understood for accommodating structured sparsity. In this article, we develop a sparse group penalization method to conduct the bi-level gene-environment (G×E) interaction study under the repeatedly measured phenotype. Within the quadratic inference function (QIF) framework, the proposed method can achieve simultaneous identification of main and interaction effects on both the group and individual level. Simulation studies have shown that the proposed method outperforms major competitors. In the case study of asthma data from the Childhood Asthma Management Program (CAMP), we conduct G×E study by using high dimensional SNP data as genetic factors and the longitudinal trait, forced expiratory volume in one second (FEV1), as the phenotype. Our method leads to improved prediction and identification of main and interaction effects with important implications.