Statistical Methods and Inference

Individual-Specific Gaussian Graphical Models for Heterogeneous Populations with Application to Epigenetic Gene Regulation in Lung Adenocarcinoma

SIREN estimates individual-specific gene regulatory networks from multi-omics data, revealing epigenetic heterogeneity in lung adenocarcinoma that correlates with patient survival.

Joint Variable Selection for Omic Biomarkers in Time-to-Event Data

The paper introduces vampW, a Bayesian method for joint variable selection in omic biomarkers and time-to-event data, showing reduced false discoveries and improved prediction.

IQC: A Novel Criterion for Assessing Feature Selection Stability in High-Dimensional Analyses

The paper introduces IQC, a metric to assess feature selection stability in high-dimensional analyses, demonstrating its effectiveness through simulations and a case study on gene expression data.

MMAD-Risk: Multivariate Mixed Survival Analysis for the Prediction of Age-Dependent Disease Risks from Plasma Proteomes

MMAD-Risk predicts age-dependent disease risks from plasma proteomes using a multivariate mixed AFT model, outperforming Cox regression on UK Biobank data.

stably: error-controlled stability selection for biomarker panel discovery

stably (Python) applies error-controlled stability selection to DIA-NN output; its 17-protein pancreatic cancer panel reached held-out AUC 0.93 versus 0.82 for a published two-protein panel rescored on the same data.

Uncovering Heterogeneous Effects via Localized Feature Selection

This paper introduces a framework for identifying disease-associated features tailored to individual profiles using localized test statistics, applied to single-cell RNA sequencing data for Alzheimer's disease.

Cooperative Learning with Penalized Linear Mixed-Effects Models for High-Dimensional Clustered Multiview Data

The paper introduces CL-pLMM, a method for integrating clustered multiview biomedical data, and shows it improves prediction in longitudinal proteomic and metabolomic studies.

Interpretable and predictive models based on high-dimensional data in ecology and evolution

The paper evaluates nine methods for variable selection and prediction in high-dimensional ecological and evolutionary data, finding that overfitting is common but prediction accuracy improves with more observations and stronger effects.

Benchmark of biomarker identification and prognostic modeling methods on diverse censored data

The paper benchmarks biomarker identification and prognostic modeling methods on censored survival data, finding that CoxBoost and Adaptive LASSO perform best across multiple metrics.

Super Learner Ensemble Modeling of CPTAC Proteomic Data for Survival Prediction in Head and Neck Squamous Cell Carcinoma

This study uses Super Learner ensemble modeling on CPTAC proteomic data to predict survival in HNSCC, finding that the ensemble method performs stably with high C-index values.

Coxmos: Interpretable survival models for high-dimensional and multi-omic data

Coxmos provides an interpretable survival analysis tool for high-dimensional and multi-omic data, outperforming existing methods while enhancing model interpretability.

Statistical knockoffs improve biomarker discovery fromtranscriptomic data

This paper applies the knockoff procedure to transcriptomic data for variable selection, finding it outperforms existing methods but is conservative in real datasets.

SuperSurv: A Unified Framework for Machine Learning Ensembles in Survival Analysis

SuperSurv provides a unified R package for building, evaluating, and interpreting ensemble survival models with right-censored data, using SHAP values and IPCW Brier risk for weighting.

A Benchmarking Study of Feature Screening Approaches Across Omics Classification Settings

This paper benchmarks feature screening methods for omics data classification, finding BcorSIS to be the most effective and efficient across various datasets.

Stochastic LASSO for extremely high-dimensional genomic data

The paper introduces Stochastic LASSO, a bootstrap-based method for feature selection in high-dimensional genomic data, showing improved performance over existing models in simulation and real cancer data.

GRASS-NB: Group-structured variable selection for spatial negative binomial data with applications to cancer registry and spatial omics

The paper introduces a Bayesian negative binomial regression model with group-structured priors for spatially correlated count data, applied to cancer registry and spatial omics datasets.

OptSurvCutR: Validated Cut-point Selection for Survival Analysis

OptSurvCutR provides a validated method for selecting optimal cut-points in survival analysis, improving accuracy over arbitrary thresholds.

Transfer Learning for Survival-based Clustering of Predictors with an Application to TP53 Mutation Annotation

The paper develops TLSCP, a transfer learning method for survival-based clustering of predictors, applied to TP53 mutation annotation, showing improved clustering and clinical interpretability.

Beyond P-values: A Multi-Metric Framework for Robust Feature Selection and Predictive Modeling

The paper introduces MIXER, a framework that combines multiple metrics for feature selection and predictive modeling, outperforming individual criteria in Alzheimer's disease analysis.

An Interpretable Deep Learning Framework for Biomarker Discovery in Complex Disease Survival Outcomes

This paper introduces SurvDNN, a deep learning framework for survival outcome modeling in complex diseases, which outperforms existing methods in biomarker identification and prediction.

Experimenting with Two Recent Feature Selection Methods for High-Dimensional Biological Data

This study compares two feature selection methods, Stabl and Nullstrap, showing Nullstrap is more accurate and efficient for high-dimensional biological data.

Reflection Knockoffs via Householder Reflection: Applications in Proteomics and Genetic Fine Mapping

This paper introduces reflection knockoffs for feature selection in highly correlated datasets, showing improved performance in proteomics and genetic fine mapping.

PLSKO: a robust knockoff generator to control false discovery rate in omics variable selection

PLSKO is a knockoff generator that controls FDR in omics variable selection, showing robustness across RNA-seq, proteomics, metabolomics, and microbiome data.

Learning Gaussian Graphical Models from Correlated Data

This paper introduces a Bootstrap algorithm to infer Gaussian Graphical Models from correlated data, showing it controls Type I error better than conventional methods in simulations and a real family-based study.

Penalized reduced rank regression for multi-outcome survival data supports a common metabolic risk score for age-related diseases

The paper introduces a penalized reduced rank regression model for multi-outcome survival data, finding a common metabolic risk score associated with age-related diseases using UK Biobank data.

Using a supervised principal components analysis for variable selection in high-dimensional datasets reduces false discovery rates

This paper introduces a supervised principal components analysis method for variable selection in high-dimensional datasets, showing it reduces false discovery rates and improves statistical inference.

NETWORK REGULARIZED ACCELERATED FAILURE TIME MODELS FOR ROBUST BIOMARKER IDENTIFICATION

This paper develops network-regularized AFT models with inverse-degree penalties for survival analysis in cancer genomics, showing improved biological interpretability and reduced false positives.

mt.surv: Multi-Threshold Survival Analysis for Associating Continuous Predictor Variables with Time-to-Event Outcomes

mt.surv is an R package that performs multi-threshold survival analysis to assess the association of continuous predictor variables with time-to-event outcomes, providing a scalar measure for ranking biological variables.

High-dimensional Biomarker Identification for Scalable and Interpretable Disease Prediction via Machine Learning Models

This paper introduces HiFIT, a framework for high-dimensional biomarker identification that improves disease prediction and feature importance detection using machine learning.

Quantification of individual dataset contributions to prediction accuracy in cooperative learning

This paper quantifies individual dataset contributions to prediction accuracy using cooperative learning in the CAMP study.