Prediction Consistency Regularization for Learning with Noise Labels Based on Contrastive Clustering.

In the classification task, label noise has a significant impact on models' performance, primarily manifested in the disruption of prediction consistency, thereby reducing the classification accuracy. This work introduces a novel prediction consistency regularization that mitigates the impact of label noise on neural networks by imposing constraints on the prediction consistency of similar samples. However, determining which samples should be similar is a primary challenge. We formalize the similar sample identification as a clustering problem and employ twin contrastive clustering (TCC) to address this issue. To ensure similarity between samples within each cluster, we enhance TCC by adjusting clustering prior to distribution using label information. Based on the adjusted TCC's clustering results, we first construct the prototype for each cluster and then formulate a prototype-based regularization term to enhance prediction consistency for the prototype within each cluster and counteract the adverse effects of label noise. We conducted comprehensive experiments using benchmark datasets to evaluate the effectiveness of our method under various scenarios with different noise rates. The results explicitly demonstrate the enhancement in classification accuracy. Subsequent analytical experiments confirm that the proposed regularization term effectively mitigates noise and that the adjusted TCC enhances the quality of similar sample recognition.

Estimation of multiple networks with common structures in heterogeneous subgroups.

Network estimation has been a critical component of high-dimensional data analysis and can provide an understanding of the underlying complex dependence structures. Among the existing studies, Gaussian graphical models have been highly popular. However, they still have limitations due to the homogeneous distribution assumption and the fact that they are only applicable to small-scale data. For example, cancers have various levels of unknown heterogeneity, and biological networks, which include thousands of molecular components, often differ across subgroups while also sharing some commonalities. In this article, we propose a new joint estimation approach for multiple networks with unknown sample heterogeneity, by decomposing the Gaussian graphical model (GGM) into a collection of sparse regression problems. A reparameterization technique and a composite minimax concave penalty are introduced to effectively accommodate the specific and common information across the networks of multiple subgroups, making the proposed estimator significantly advancing from the existing heterogeneity network analysis based on the regularized likelihood of GGM directly and enjoying scale-invariant, tuning-insensitive, and optimization convexity properties. The proposed analysis can be effectively realized using parallel computing. The estimation and selection consistency properties are rigorously established. The proposed approach allows the theoretical studies to focus on independent network estimation only and has the significant advantage of being both theoretically and computationally applicable to large-scale data. Extensive numerical experiments with simulated data and the TCGA breast cancer data demonstrate the prominent performance of the proposed approach in both subgroup and network identifications.

Endocrine disrupting chemical mixture exposure and risk of papillary thyroid cancer in U.S. military personnel: A nested case-control study.

Single-pollutant methods to evaluate associations between endocrine disrupting chemicals (EDCs) and thyroid cancer risk may not reflect realistic human exposures. Therefore, we evaluated associations between exposure to a mixture of 18 EDCs, including polychlorinated biphenyls (PCBs), brominated flame retardants, and organochlorine pesticides, and risk of papillary thyroid cancer (PTC), the most common thyroid cancer histological subtype. We conducted a nested case-control study among U.S. military servicemembers of 652 histologically-confirmed PTC cases diagnosed between 2000 and 2013 and 652 controls, matched on birth year, sex, race/ethnicity, military component (active duty/reserve), and serum sample timing. We estimated mixture odds ratios (OR), 95% confidence intervals (95% CI), and standard errors (SE) for associations between pre-diagnostic serum EDC mixture concentrations, overall PTC risk, and risk of histological subtypes of PTC (classical, follicular), adjusted for body mass index and military branch, using quantile g-computation. Additionally, we identified relative contributions of individual mixture components to PTC risk, represented by positive and negative weights (w). A one-quartile increase in the serum mixture concentration was associated with a non-statistically significant increase in overall PTC risk (OR = 1.19; 95% CI = 0.91, 1.56; SE = 0.14). Stratified by histological subtype and race (White, Black), a one-quartile increase in the mixture was associated with increased classical PTC risk among those of White race (OR = 1.59; 95% CI = 1.06, 2.40; SE = 0.21), but not of Black race (OR = 0.95; 95% CI = 0.34, 2.68; SE = 0.53). PCBs 180, 199, and 118 had the greatest positive weights driving this association among those of White race (w = 0.312, 0.255, and 0.119, respectively). Findings suggest that exposure to an EDC mixture may be associated with increased classical PTC risk. These findings warrant further investigation in other study populations to better understand PTC risk by histological subtype and race.

Information-incorporated sparse hierarchical cancer heterogeneity analysis.

Cancer heterogeneity analysis is essential for precision medicine. Most of the existing heterogeneity analyses only consider a single type of data and ignore the possible sparsity of important features. In cancer clinical practice, it has been suggested that two types of data, pathological imaging and omics data, are commonly collected and can produce hierarchical heterogeneous structures, in which the refined sub-subgroup structure determined by omics features can be nested in the rough subgroup structure determined by the imaging features. Moreover, sparsity pursuit has extraordinary significance and is more challenging for heterogeneity analysis, because the important features may not be the same in different subgroups, which is ignored by the existing heterogeneity analyses. Fortunately, rich information from previous literature (for example, those deposited in PubMed) can be used to assist feature selection in the present study. Advancing from the existing analyses, in this study, we propose a novel sparse hierarchical heterogeneity analysis framework, which can integrate two types of features and incorporate prior knowledge to improve feature selection. The proposed approach has satisfactory statistical properties and competitive numerical performance. A TCGA real data analysis demonstrates the practical value of our approach in analyzing data heterogeneity and sparsity.

Organochlorine pesticides and risk of papillary thyroid cancer in U.S. military personnel: a nested case-control study.

BACKGROUND: The effects of organochlorine pesticide (OCP) exposure on the development of human papillary thyroid cancer (PTC) are not well understood. A nested case-control study was conducted with data from the U.S. Department of Defense Serum Repository (DoDSR) cohort between 2000 and 2013 to assess associations of individual OCPs serum concentrations with PTC risk. METHODS: This study included 742 histologically confirmed PTC cases (341 females, 401 males) and 742 individually-matched controls with pre-diagnostic serum samples selected from the DoDSR. Associations between categories of lipid-corrected serum concentrations of seven OCPs and PTC risk were evaluated for classical PTC and follicular PTC using conditional logistic regression, adjusted for body mass index category and military branch to compute odds ratios (OR) and 95% confidence intervals (CIs). Effect modification by sex, birth cohort, and race was examined. RESULTS: There was no evidence of associations between most of the OCPs and PTC, overall or stratified by histological subtype. Overall, there was no evidence of an association between hexachlorobenzene (HCB) and PTC, but stratified by histological subtype HCB was associated with significantly increased risk of classical PTC (third tertile above the limit of detection (LOD) vs.

Neurogenetic underpinnings of nicotine use severity: Integrating the brain transcriptomes and GWAS variants via network approaches.

Our study focused on human brain transcriptomes and the genetic risks of cigarettes per day (CPD) to investigate the neurogenetic mechanisms of individual variation in nicotine use severity. We constructed whole-brain and intramodular region-specific coexpression networks using BrainSpan's transcriptomes, and the genomewide association studies identified risk variants of CPD, confirmed the associations between CPD and each gene set in the region-specific subnetworks using an independent dataset, and conducted bioinformatic analyses. Eight brain-region-specific coexpression subnetworks were identified in association with CPD: amygdala, hippocampus, medial prefrontal cortex (MPFC), orbitofrontal cortex (OPFC), dorsolateral prefrontal cortex, striatum, mediodorsal nucleus of the thalamus (MDTHAL), and primary motor cortex (M1C). Each gene set in the eight subnetworks was associated with CPD. We also identified three hub proteins encoded by GRIN2A in the amygdala, PMCA2 in the hippocampus, MPFC, OPFC, striatum, and MDTHAL, and SV2B in M1C. Intriguingly, the pancreatic secretion pathway appeared in all the significant protein interaction subnetworks, suggesting pleiotropic effects between cigarette smoking and pancreatic diseases. The three hub proteins and genes are implicated in stress response, drug memory, calcium homeostasis, and inhibitory control. These findings provide novel evidence of the neurogenetic underpinnings of smoking severity.

Hierarchical False Discovery Rate Control for High-dimensional Survival Analysis with Interactions.

With the development of data collection techniques, analysis with a survival response and high-dimensional covariates has become routine. Here we consider an interaction model, which includes a set of low-dimensional covariates, a set of high-dimensional covariates, and their interactions. This model has been motivated by gene-environment (G-E) interaction analysis, where the E variables have a low dimension, and the G variables have a high dimension. For such a model, there has been extensive research on estimation and variable selection. Comparatively, inference studies with a valid false discovery rate (FDR) control have been very limited. The existing high-dimensional inference tools cannot be directly applied to interaction models, as interactions and main effects are not "equal". In this article, for high-dimensional survival analysis with interactions, we model survival using the Accelerated Failure Time (AFT) model and adopt a "weighted least squares + debiased Lasso" approach for estimation and selection. A hierarchical FDR control approach is developed for inference and respect of the "main effects, interactions" hierarchy. The asymptotic distribution properties of the debiased Lasso estimators are rigorously established. Simulation demonstrates the satisfactory performance of the proposed approach, and the analysis of a breast cancer dataset further establishes its practical utility.

FunctanSNP: an R package for functional analysis of dense SNP data (with interactions).

SUMMARY: Densely measured SNP data are routinely analyzed but face challenges due to its high dimensionality, especially when gene-environment interactions are incorporated. In recent literature, a functional analysis strategy has been developed, which treats dense SNP measurements as a realization of a genetic function and can 'bypass' the dimensionality challenge. However, there is a lack of portable and friendly software, which hinders practical utilization of these functional methods. We fill this knowledge gap and develop the R package FunctanSNP. This comprehensive package encompasses estimation, identification, and visualization tools and has undergone extensive testing using both simulated and real data, confirming its reliability. FunctanSNP can serve as a convenient and reliable tool for analyzing SNP and other densely measured data. AVAILABILITY AND IMPLEMENTATION: The package is available at https://CRAN.R-project.org/package=FunctanSNP.

HETEROGENEITY ANALYSIS VIA INTEGRATING MULTI-SOURCES HIGH-DIMENSIONAL DATA WITH APPLICATIONS TO CANCER STUDIES.

This study has been motivated by cancer research, in which heterogeneity analysis plays an important role and can be roughly classified as unsupervised or supervised. In supervised heterogeneity analysis, the finite mixture of regression (FMR) technique is used extensively, under which the covariates affect the response differently in subgroups. High-dimensional molecular and, very recently, histopathological imaging features have been analyzed separately and shown to be effective for heterogeneity analysis. For simpler analysis, they have been shown to contain overlapping, but also independent information. In this article, our goal is to conduct the first and more effective FMR-based cancer heterogeneity analysis by integrating high-dimensional molecular and histopathological imaging features. A penalization approach is developed to regularize estimation, select relevant variables, and, equally importantly, promote the identification of independent information. Consistency properties are rigorously established. An effective computational algorithm is developed. A simulation and an analysis of The Cancer Genome Atlas (TCGA) lung cancer data demonstrate the practical effectiveness of the proposed approach. Overall, this study provides a practical and useful new way of conducting supervised cancer heterogeneity analysis.

A General Framework for Identifying Hierarchical Interactions and Its Application to Genomics Data.

The analysis of hierarchical interactions has long been a challenging problem due to the large number of candidate main effects and interaction effects, and the need for accommodating the "main effects, interactions" hierarchy. The two-stage analysis methods enjoy simplicity and low computational cost, but contradict the fact that the outcome of interest is attributable to the joint effects of multiple main factors and their interactions. The existing joint analysis methods can accurately describe the underlying data generating process, but suffer from prohibitively high computational cost. And it is not straightforward to extend their optimization algorithms to general loss functions. To address this need, we develop a new computational method that is much faster than the existing joint analysis methods and rivals the runtimes of two-stage analysis. The proposed method, HierFabs, adopts the framework of the forward and backward stagewise algorithm and enjoys computational efficiency and broad applicability. To accommodate hierarchy without imposing additional constraints, it has newly developed forward and backward steps. It naturally accommodates the strong and weak hierarchy, and makes optimization much simpler and faster than in the existing studies. Optimality of HierFabs sequences is investigated theoretically. Simulations show that it outperforms the existing methods. The analysis of TCGA data on melanoma demonstrates its competitive practical performance.

Comparative Effectiveness Analysis of Lumpectomy and Mastectomy for Elderly Female Breast Cancer Patients: A Deep Learning-based Big Data Analysis.

Objectives: To evaluate the comparative effectiveness of treatments, a randomized clinical trial remains the gold standard but can be challenged by a high cost, a limited sample size, an inability to fully reflect the real world, and feasibility concerns. The objective is to showcase a big data approach that takes advantage of large electronic medical record (EMR) data to emulate clinical trials. To overcome the limitations of regression analysis, a deep learning-based analysis pipeline was developed. Study Design and Setting: Lumpectomy (breast-conserving surgery) and mastectomy are the two most commonly used surgical procedures for early-stage female breast cancer patients. An emulation trial was designed using the Surveillance, Epidemiology, and End Results (SEER)-Medicare data to evaluate their relative effectiveness in overall survival. The analysis pipeline consisted of a propensity score step, a weighted survival analysis step, and a bootstrap inference step. Results: A total of 65,997 subjects were enrolled in the emulated trial, with 50,704 and 15,293 in the lumpectomy and mastectomy arms, respectively. The two surgery procedures had comparable effects in terms of overall survival (survival year change = 0.08, 95% confidence interval (CI): -0.08, 0.25) for the elderly SEER-Medicare early-stage female breast cancer patients. Conclusion: This study demonstrated the power of "mining large EMR data + deep learning-based analysis," and the proposed analysis strategy and technique can be potentially broadly applicable. It provided convincing evidence of the comparative effectiveness of lumpectomy and mastectomy.

ZINBMM: a general mixture model for simultaneous clustering and gene selection using single-cell transcriptomic data.

Clustering is a critical component of single-cell RNA sequencing (scRNA-seq) data analysis and can help reveal cell types and infer cell lineages. Despite considerable successes, there are few methods tailored to investigating cluster-specific genes contributing to cell heterogeneity, which can promote biological understanding of cell heterogeneity. In this study, we propose a zero-inflated negative binomial mixture model (ZINBMM) that simultaneously achieves effective scRNA-seq data clustering and gene selection. ZINBMM conducts a systemic analysis on raw counts, accommodating both batch effects and dropout events. Simulations and the analysis of five scRNA-seq datasets demonstrate the practical applicability of ZINBMM.

Prior information-assisted integrative analysis of multiple datasets.

MOTIVATION: Analyzing genetic data to identify markers and construct predictive models is of great interest in biomedical research. However, limited by cost and sample availability, genetic studies often suffer from the "small sample size, high dimensionality" problem. To tackle this problem, an integrative analysis that collectively analyzes multiple datasets with compatible designs is often conducted. For regularizing estimation and selecting relevant variables, penalization and other regularization techniques are routinely adopted. "Blindly" searching over a vast number of variables may not be efficient. RESULTS: We propose incorporating prior information to assist integrative analysis of multiple genetic datasets. To obtain accurate prior information, we adopt a convolutional neural network with an active learning strategy to label textual information from previous studies. Then the extracted prior information is incorporated using a group LASSO-based technique. We conducted a series of simulation studies that demonstrated the satisfactory performance of the proposed method. Finally, data on skin cutaneous melanoma are analyzed to establish practical utility. AVAILABILITY AND IMPLEMENTATION: Code is available at https://github.com/ldz7/PAIA. The data that support the findings in this article are openly available in TCGA (The Cancer Genome Atlas) at https://portal.gdc.cancer.gov/.

Aligned deep neural network for integrative analysis with high-dimensional input.

OBJECTIVE: Deep neural network (DNN) techniques have demonstrated significant advantages over regression and some other techniques. In recent studies, DNN-based analysis has been conducted on data with high-dimensional input such as omics measurements. In such analysis, regularization, in particular penalization, has been applied to regularize estimation and distinguish relevant input variables from irrelevant ones. A unique challenge arises from the "lack of information" attributable to high dimensionality of input and limited size of training data. For many data/studies, there exist other data/studies that may be relevant and can potentially provide additional information to boost performance. METHODS: In this study, we conduct integrative analysis of multiple independent datasets/studies, with the goal of borrowing information across each other and improving overall performance. Significantly different from regression-based integrative analysis (where alignment can be easily achieved based on covariates), alignment across multiple DNNs can be nontrivial. We develop ANNI, an Aligned DNN technique for Integrative analysis with high-dimensional input. Penalization is applied for regularized estimation, selection of important input variables, and, equally importantly, information borrowing across multiple DNNs. An effective computational algorithm is developed. RESULTS: Extensive simulations demonstrate competitive performance of the proposed technique. The analysis of cancer omics data further establishes its practical utility.

Pathological imaging-assisted cancer gene-environment interaction analysis.

Gene-environment (G-E) interactions have important implications for cancer outcomes and phenotypes beyond the main G and E effects. Compared to main-effect-only analysis, G-E interaction analysis more seriously suffers from a lack of information caused by higher dimensionality, weaker signals, and other factors. It is also uniquely challenged by the "main effects, interactions" variable selection hierarchy. Effort has been made to bring in additional information to assist cancer G-E interaction analysis. In this study, we take a strategy different from the existing literature and borrow information from pathological imaging data. Such data are a "byproduct" of biopsy, enjoys broad availability and low cost, and has been shown as informative for modeling prognosis and other cancer outcomes/phenotypes in recent studies. Building on penalization, we develop an assisted estimation and variable selection approach for G-E interaction analysis. The approach is intuitive, can be effectively realized, and has competitive performance in simulation. We further analyze The Cancer Genome Atlas (TCGA) data on lung adenocarcinoma (LUAD). The outcome of interest is overall survival, and for G variables, we analyze gene expressions. Assisted by pathological imaging data, our G-E interaction analysis leads to different findings with competitive prediction performance and stability.

Bi-level structured functional analysis for genome-wide association studies.

Genome-wide association studies (GWAS) have led to great successes in identifying genotype-phenotype associations for complex human diseases. In such studies, the high dimensionality of single nucleotide polymorphisms (SNPs) often makes analysis difficult. Functional analysis, which interprets SNPs densely distributed in a chromosomal region as a continuous process rather than discrete observations, has emerged as a promising avenue for overcoming the high dimensionality challenges. However, the majority of the existing functional studies continue to be individual SNP based and are unable to sufficiently account for the intricate underpinning structures of SNP data. SNPs are often found in groups (e.g., genes or pathways) and have a natural group structure. Additionally, these SNP groups can be highly correlated with coordinated biological functions and interact in a network. Motivated by these unique characteristics of SNP data, we develop a novel bi-level structured functional analysis method and investigate disease-associated genetic variants at the SNP level and SNP group level simultaneously. The penalization technique is adopted for bi-level selection and also to accommodate the group-level network structure. Both the estimation and selection consistency properties are rigorously established. The superiority of the proposed method over alternatives is shown through extensive simulation studies. A type 2 diabetes SNP data application yields some biologically intriguing results.

Bayesian finite mixture of regression analysis for cancer based on histopathological imaging-environment interactions.

Cancer is a heterogeneous disease. Finite mixture of regression (FMR)-as an important heterogeneity analysis technique when an outcome variable is present-has been extensively employed in cancer research, revealing important differences in the associations between a cancer outcome/phenotype and covariates. Cancer FMR analysis has been based on clinical, demographic, and omics variables. A relatively recent and alternative source of data comes from histopathological images. Histopathological images have been long used for cancer diagnosis and staging. Recently, it has been shown that high-dimensional histopathological image features, which are extracted using automated digital image processing pipelines, are effective for modeling cancer outcomes/phenotypes. Histopathological imaging-environment interaction analysis has been further developed to expand the scope of cancer modeling and histopathological imaging-based analysis. Motivated by the significance of cancer FMR analysis and a still strong demand for more effective methods, in this article, we take the natural next step and conduct cancer FMR analysis based on models that incorporate low-dimensional clinical/demographic/environmental variables, high-dimensional imaging features, as well as their interactions. Complementary to many of the existing studies, we develop a Bayesian approach for accommodating high dimensionality, screening out noises, identifying signals, and respecting the "main effects, interactions" variable selection hierarchy. An effective computational algorithm is developed, and simulation shows advantageous performance of the proposed approach. The analysis of The Cancer Genome Atlas data on lung squamous cell cancer leads to interesting findings different from the alternative approaches.

Unified model-free interaction screening via CV-entropy filter.

For many practical high-dimensional problems, interactions have been increasingly found to play important roles beyond main effects. A representative example is gene-gene interaction. Joint analysis, which analyzes all interactions and main effects in a single model, can be seriously challenged by high dimensionality. For high-dimensional data analysis in general, marginal screening has been established as effective for reducing computational cost, increasing stability, and improving estimation/selection performance. Most of the existing marginal screening methods are designed for the analysis of main effects only. The existing screening methods for interaction analysis are often limited by making stringent model assumptions, lacking robustness, and/or requiring predictors to be continuous (and hence lacking flexibility). A unified marginal screening approach tailored to interaction analysis is developed, which can be applied to regression, classification, and survival analysis. Predictors are allowed to be continuous and discrete. The proposed approach is built on Coefficient of Variation (CV) filters based on information entropy. Statistical properties are rigorously established. It is shown that the CV filters are almost insensitive to the distribution tails of predictors, correlation structure among predictors, and sparsity level of signals. An efficient two-stage algorithm is developed to make the proposed approach scalable to ultrahigh-dimensional data. Simulations and the analysis of TCGA LUAD data further establish the practical superiority of the proposed approach.

SARS-CoV-2 mRNA vaccines decouple anti-viral immunity from humoral autoimmunity.

mRNA-based vaccines dramatically reduce the occurrence and severity of COVID-19, but are associated with rare vaccine-related adverse effects. These toxicities, coupled with observations that SARS-CoV-2 infection is associated with autoantibody development, raise questions whether COVID-19 vaccines may also promote the development of autoantibodies, particularly in autoimmune patients. Here we used Rapid Extracellular Antigen Profiling to characterize self- and viral-directed humoral responses after SARS-CoV-2 mRNA vaccination in 145 healthy individuals, 38 patients with autoimmune diseases, and 8 patients with mRNA vaccine-associated myocarditis. We confirm that most individuals generated robust virus-specific antibody responses post vaccination, but that the quality of this response is impaired in autoimmune patients on certain modes of immunosuppression. Autoantibody dynamics are remarkably stable in all vaccinated patients compared to COVID-19 patients that exhibit an increased prevalence of new autoantibody reactivities. Patients with vaccine-associated myocarditis do not have increased autoantibody reactivities relative to controls. In summary, our findings indicate that mRNA vaccines decouple SARS-CoV-2 immunity from autoantibody responses observed during acute COVID-19.