Selected Publications

 

Univariate-Guided Sparse Regression for Biobank-Scale High-Dimensional Omics Data

snpnet
Lasso
High-dimensional regression
penalized regression

We propose a univariate-guided sparse regression framework for efficient analysis of biobank-scale high-dimensional omics data, combining marginal screening with penalized regression.

Nov 28, 2025
Joshua Richland, Tuomo Kiiskinen, William Wang, Sophia Lu, Balasubramanian Narasimhan, Trevor Hastie, Manuel Rivas, Robert Tibshirani
 

Pretraining and the Lasso

Lasso
penalized regression
High-dimensional regression
transfer learning

We study the use of pretraining data to improve lasso regression, providing theoretical and empirical results on when and how auxiliary data can boost prediction and variable selection.

Aug 1, 2025
Erin Craig, Mert Pilanci, Thomas Le Menestrel, Balasubramanian Narasimhan, Manuel A. Rivas, Stein-Erik Gullaksen, Roozbeh Dehghannasiri, Julia Salzman, Jonathan Taylor, Robert Tibshirani
 

Cooperative Learning for Multiview Analysis

snpnet
Lasso
High-dimensional regression
penalized regression

We propose a method for supervised learning with multiple sets of features (“views”). The multiview problem is especially important in biology and medicine, where “-omics” data, such as genomics, proteomics, and radiomics, are measured on a common set of samples.

Sep 12, 2022
Daisy Ding, Shuangning Li, Balasubramanian Narasimhan, Robert Tibshirani
 

Penalized Regression for Left-Truncated and Right-Censored Survival Data

penalized regression
survival analysis

We propose penalized regression methods for left-truncated and right-censored survival data, extending the elastic net penalty to handle delayed entry in Cox proportional hazards models.

Nov 10, 2021
Sarah F. McGough, Devin Incerti, Svetlana Lyalina, Ryan Copping, Balasubramanian Narasimhan, Robert Tibshirani
 

Fast Numerical Optimization for Genome Sequencing Data in Population Biobanks

snpnet
Lasso
High-dimensional regression
penalized regression

We develop two efficient solvers for optimization problems arising from large-scale regularized regressions on millions of genetic variants sequenced from hundreds of thousands of individuals.

Jun 21, 2021
Ruilin Li, Christopher Chang, Yosuke Tanigawa, Balasubramanian Narasimhan, Trevor Hastie, Robert Tibshirani, Manuel Rivas
 

Elastic Net Regularization Paths for All Generalized Linear Models

glmnet
Lasso
High-dimensional regression
penalized regression

This paper describes some of the substantial enhancements to the glmnet R package ver 4.1+. All programmed GLM families are accommodated through a family() argument. We also discuss relaxed fits, and facilities for modeling stop/start data and strata in survival models.

Mar 5, 2021
J. Kenneth Tay, Balasubramanian Narasimhan, Trevor Hastie
 

Evaluation of Individual and Ensemble Probabilistic Forecasts of COVID-19 Mortality in the US

COVID-19
COVIDCAST
forecasting

Short-term probabilistic forecasts of the trajectory of the COVID-19 pandemic in the United States have served as a visible and important communication channel between the scientific modeling community and both the general public and decision-makers.

Feb 5, 2021
EY Cramer, et al
 

Defining the Features and Duration of Antibody Responses to SARS-CoV-2 Infection Associated with Disease Severity and Outcome

SARS-COV-2

SARS-CoV-2-specific antibodies, particularly those preventing viral spike receptor binding domain (RBD) interaction with host angiotensin-converting enzyme 2 (ACE2) receptor, can neutralize the virus.

Dec 7, 2020
Katharina Röltgen, Abigail E Powell, Oliver F Wirz, Bryan A Stevens, Catherine A Hogan, Javaria Najeeb, Molly Hunter, Hannah Wang, Malaya K Sahoo, ChunHong Huang, Fumiko Yamamoto, Monali Manohar, Justin Manalac, Ana R Otrelo-Cardoso, Tho D Pham, Arjun Rustagi, Angela J Rogers, Nigam H Shah, Catherine A Blish, Jennifer R Cochran, Theodore S Jardetzky, James L Zehnder, Taia T Wang, Balasubramanian Narasimhan, Saurabh Gombar, Robert Tibshirani, Kari C Nadeau, Peter S Kim, Benjamin A Pinsky, Scott D Boyd
 

CVXR: An R Package for Disciplined Convex Optimization

convex optimization
R

CVXR is an R package that provides an object-oriented modeling language for convex optimization, similar to CVX, CVXPY, and YALMIP. Users formulate problems in natural mathematical syntax by combining constants, variables, and parameters with a library of functions with known mathematical properties. CVXR then verifies convexity via signed disciplined convex programming, converts to standard conic form, and passes the problem to cone solvers such as ECOS or SCS.

Sep 13, 2020
Anqi Fu, Balasubramanian Narasimhan, Stephen Boyd
 

A g-Modeling Program for Deconvolution and Empirical Bayes Estimation

Empirical Bayes
deconvolution

Empirical Bayes inference assumes an unknown prior density g has yielded unobservable parameters, each producing an independent observation. The core challenge involves recovering the prior distribution from observed data through deconvolution. We propose restricting the prior to a parametric exponential family to make estimation tractable, and provide the R package deconvolveR implementing practical Bayes deconvolution and g-modeling techniques.

Sep 2, 2020
Balasubramanian Narasimhan, Bradley Efron
 

Derivation and Validation of a Prognostic Score for Neonatal Mortality in Ethiopia

clinical
global health

A case-control study developing and validating a prognostic score for neonatal mortality in a resource-limited setting in Ethiopia.

May 18, 2020
Rishi Mediratta, Abebe T. Amare, Roop Behl, Bradley Efron, Balasubramanian Narasimhan, Alemayehu Teklu, Abere Shehibo, Mamo Ayalew, Satyan Kache
 

The Automatic Construction of Bootstrap Confidence Intervals

Bootstrap
confidence intervals

We consider the problem of setting approximate confidence intervals for a single parameter in a multiparameter family. A new set of algorithms automates the construction of bootstrap confidence intervals, substituting computer power for the need to individually program particular applications. The BCA (bias-corrected and accelerated) bootstrap method provides second-order accuracy, an order of magnitude improvement over standard intervals.

Mar 12, 2020
Bradley Efron, Balasubramanian Narasimhan
 

Components of Genetic Associations across 2,138 Phenotypes in the UK Biobank Highlight Novel Adipocyte Biology

genomics
UK Biobank

We apply truncated singular value decomposition (DeGAs) to matrices of summary statistics derived from genome-wide association analyses across 2,138 phenotypes in 337,199 White British individuals in the UK Biobank. This decomposition reveals latent components driving genetic associations, identifying variants in GPR151 and PDE3B that substantially contribute to obesity-related traits, validated through functional studies in adipocyte biology.

Sep 6, 2019
Yosuke Tanigawa, Jiehan Li, Johanne Marie Justesen, Heiko Horn, Matthew Aguirre, Christopher DeBoever, Christopher Chang, Balasubramanian Narasimhan, Kasper Lage, Trevor John Hastie, Chong Yon Park, Gill Bejerano, Erik Ingelsson, Manuel Rivas
 

Imputation of Mixed Data with Multilevel Singular Value Decomposition

missing data
imputation
SVD

We present a single imputation method for multilevel data containing mixed quantitative and qualitative variables with missing values. The approach is based on multilevel singular value decomposition, which decomposes data variability into between- and within-group components and applies SVD to each. The method is computationally efficient, handles diverse data types, and is demonstrated on simulations and medical data from multiple hospitals.

May 28, 2019
François Husson, Julie Josse, Balasubramanian Narasimhan, Geneviève Robin
 

A Scalable Discrete-Time Survival Model for Neural Networks

There is currently great interest in applying neural networks to prediction tasks in medicine. It is important for predictive models to be able to use survival data, where each patient has a known follow-up time and event/censoring indicator.

Jan 25, 2019
Michael Gensheimer, Balasubramanian Narasimhan
 

Big Data Modeling to Predict Platelet Usage and Minimize Wastage in a Tertiary Care System

clinical
prediction

We develop a statistical model using hospital patient data to quantitatively forecast platelet transfusion needs days in advance. Analyzing 29 months of data from Stanford Hospital, the model reduces the platelet expiration rate from 10.5% to 3.2% while maintaining adequate inventory. Nationwide implementation could save approximately $80 million annually.

Oct 17, 2017
Leying Guan, Xiaoying Tian, Saurabh Gombar, Allison J. Zemek, Gomathi Krishnan, Robert Scott, Balasubramanian Narasimhan, Robert J. Tibshirani, Tho D. Pham
No matching items

Interviews

An Interview with Bradley Efron Balasubramanian Narasimhan — International Statistical Review, 2020

PDF