Model-free prediction test with application to genomics data.

Testing the significance of predictors in a regression model is one of the most important topics in statistics. This problem is especially difficult without any parametric assumptions on the data. This paper aims to test the null hypothesis that given confounding variables Z, X does not significantly contribute to the prediction of Y under the model-free setting, where X and Z are possibly high dimensional. We propose a general framework that first fits nonparametric machine learning regression algorithms on [Formula: see text] and [Formula: see text], then compares the prediction power of the two models. The proposed method allows us to leverage the strength of the most powerful regression algorithms developed in the modern machine learning community. The P value for the test can be easily obtained by permutation. In simulations, we find that the proposed method is more powerful compared to existing methods. The proposed method allows us to draw biologically meaningful conclusions from two gene expression data analyses without strong distributional assumptions: 1) testing the prediction power of sequencing RNA for the proteins in cellular indexing of transcriptomes and epitopes by sequencing data and 2) identification of spatially variable genes in spatially resolved transcriptomics data.

Robust probabilistic modeling for single-cell multimodal mosaic integration and imputation via scVAEIT.

Recent advances in single-cell technologies enable joint profiling of multiple omics. These profiles can reveal the complex interplay of different regulatory layers in single cells; still, new challenges arise when integrating datasets with some features shared across experiments and others exclusive to a single source; combining information across these sources is called mosaic integration. The difficulties lie in imputing missing molecular layers to build a self-consistent atlas, finding a common latent space, and transferring learning to new data sources robustly. Existing mosaic integration approaches based on matrix factorization cannot efficiently adapt to nonlinear embeddings for the latent cell space and are not designed for accurate imputation of missing molecular layers. By contrast, we propose a probabilistic variational autoencoder model, scVAEIT, to integrate and impute multimodal datasets with mosaic measurements. A key advance is the use of a missing mask for learning the conditional distribution of unobserved modalities and features, which makes scVAEIT flexible to combine different panels of measurements from multimodal datasets accurately and in an end-to-end manner. Imputing the masked features serves as a supervised learning procedure while preventing overfitting by regularization. Focusing on gene expression, protein abundance, and chromatin accessibility, we validate that scVAEIT robustly imputes the missing modalities and features of cells biologically different from the training data. scVAEIT also adjusts for batch effects while maintaining the biological variation, which provides better latent representations for the integrated datasets. We demonstrate that scVAEIT significantly improves integration and imputation across unseen cell types, different technologies, and different tissues.

Subtypes of Dual Users of Combustible and Electronic Cigarettes: Longitudinal Changes in Product Use and Dependence Symptomatology.

INTRODUCTION: Cross-sectional surveys found behavioral heterogeneity among dual users of combustible and electronic cigarettes. Yet, prior classification did not reflect dynamic interactions between cigarette and e-cigarette consumption, which may reveal changes in product-specific dependence. The contexts of dual use that could inform intervention were also understudied. METHODS: This study conducted secondary analysis on 13 waves of data from 227 dual users who participated in a 2-year observational study. The k-means method for joint trajectories of cigarette and e-cigarette consumption was adopted to identify the subtypes of dual users. The time-varying effect model was used to characterize the subtype-specific trajectories of cigarette and e-cigarette dependence. The subtypes were also compared in terms of use contexts. RESULTS: The four clusters were identified: light dual users, predominant vapers, heavy dual users, and predominant smokers. Although heavy dual users and predominant smokers both smoked heavily at baseline, by maintaining vaping at the weekly to daily level the heavy dual users were able to considerably reduce cigarette use. Yet, the heavy dual users' drop in cigarette dependence was not as dramatic as their drop in cigarette consumption. Predominant vapers appeared to engage in substitution, as they decreased their smoking and increased their e-cigarette dependence. They were also more likely to live in environments with smoking restrictions and report that their use of e-cigarettes reduced cigarette craving and smoking frequency. CONCLUSIONS: Environmental constraints can drive substitution behavior and the substitution behavior is able to be sustained if people find the substitute to be effective. IMPLICATIONS: This study characterizes subtypes of dual users based on the dynamic interactions between cigarette use and e-cigarette use as well as product-specific trajectories of dependence. The subtypes differ in not only sociodemographic characteristics but also contexts of cigarette and e-cigarette use. Higher motivation to use e-cigarettes to quit smoking and less permissive environment for smoking may promote substitution of cigarettes by e-cigarettes.

Causal discoveries for high dimensional mixed data.

Causal relationships are of crucial importance for biological and medical research. Algorithms have been proposed for causal structure learning with graphical visualizations. While much of the literature focuses on biological studies where data often follow the same distribution, for example, the normal distribution for all variables, challenges emerge from epidemiological and clinical studies where data are often mixed with continuous, binary, and ordinal variables. We propose to use a mixed latent Gaussian copula model to estimate the underlying correlation structure via the rank correlation for mixed data. This correlation structure is then incorporated into a popular causal discovery algorithm, the PC algorithm, to identify causal structures. The proposed algorithm, called the latent-PC algorithm, is able to discover the true causal structure consistently under mild conditions in high dimensional settings. From simulation studies, the latent-PC algorithm delivers a competitive performance in terms of a similar or higher true positive rate and a similar or lower false positive rate, compared with other variants of the PC algorithm. In the high dimensional settings where the number of variables is more than the number of observations, the causal graphs identified by the latent-PC algorithm are closer to the true causal structures, compared to other competing algorithms. Further, we demonstrate the utility of the latent-PC algorithm in a real dataset for hepatocellular carcinoma. Causal structures for patient survival are visualized and connected with clinical interpretations in the literature.

Validating E-Cigarette Dependence Scales Based on Dynamic Patterns of Vaping Behaviors.

INTRODUCTION: Existing e-cigarette dependence scales are mainly validated based on retrospective overall consumption or perception. Further, given that the majority of adult e-cigarette users also use combustible cigarettes, it is important to determine whether e-cigarette dependence scales capture the product-specific dependence. This study fills in the current knowledge gaps by validating e-cigarette dependence scales using novel indices of dynamic patterns of e-cigarette use behaviors and examining the association between dynamic patterns of smoking and e-cigarette dependence among dual users. METHODS: Secondary analysis was conducted on the 2-week ecological momentary assessment data from 116 dual users. The Smoothly Clipped Absolute Deviation penalty (SCAD) was adopted to select important indices for dynamic patterns of consumption or craving and estimate their associations with e-cigarette dependence scales. RESULTS: The fitted linear regression models support the hypothesis that higher e-cigarette dependence is associated with higher levels of e-cigarette consumption and craving as well as lower instability of e-cigarette consumption. Controlling for dynamic patterns of vaping, dual users with lower e-cigarette dependence tend to report higher day-to-day dramatic changes in combustible cigarette consumption but not higher average levels of smoking. CONCLUSIONS: We found that more stable use patterns are related to higher levels of dependence, which has been demonstrated in combustible cigarettes and we have now illustrated in e-cigarettes. Furthermore, the e-cigarette dependence scales may capture the product-specific average consumption but not product-specific instability of consumption. IMPLICATIONS: This study provides empirical support for three e-cigarette dependence measures: PS-ECDI, e-FTCD, and e-WISDM, based on dynamic patterns of e-cigarette consumption and craving revealed by EMA data that have great ecological validity. This is the first study that introduces novel indices of dynamic patterns and demonstrates their potential applications in vaping research.

The association between short-term emotion dynamics and cigarette dependence: A comprehensive examination of dynamic measures.

BACKGROUND: The association between short-term emotion dynamics and long-term psychopathology has been well established in the psychology literature. Yet, dynamic measures for inertia and instability of negative and positive affect have not been studied in terms of their association with cigarette dependence. This study builds an important bridge between the psychology and substance use literatures by introducing these novel measures and conducting a comprehensive examination of such association with intervention implications. METHODS: This study conducted secondary analysis on the data from a community sample of 136 dual users (e-cigarette + cigarette) and 101 exclusive smokers who completed both the two-week ecological momentary assessment (EMA) and cigarette dependence assessments in a recent study. RESULTS: Among dual users, a higher average level of negative affect, lower inertia of negative affect (i.e., less sustained negative affect), and higher instability of positive affect (i.e., greater magnitude of changes in positive affect) were associated with higher cigarette dependence. The patterns of associations among exclusive smokers were, however, different. Higher inertia of negative affect, lower instability of positive affect, and higher variability of negative affect were associated with higher dependence. CONCLUSIONS: The results illustrate the importance of examining not only negative affect but also positive affect in order to fully understand the association between emotion dynamics and cigarette dependence. The different patterns of association between emotion dynamics and cigarette dependence across the two groups of cigarette users also call for future research that is designed to compare cigarettes and e-cigarettes in terms of their effects on emotion regulation.