Structured, sparse regression with application to HIV drug resistance

TL;DR

This paper presents a structured, sparse regression method tailored for predicting HIV drug resistance, enhancing interpretability while maintaining predictive accuracy, with theoretical insights and simulation validation.

Contribution

A novel forward stepwise regression approach that incorporates predictor structure, improving interpretability in HIV drug resistance prediction.

Findings

Improved interpretability of drug resistance models

Comparable predictive accuracy to standard methods

Theoretical analysis and simulation validation

Abstract

We introduce a new version of forward stepwise regression. Our modification finds solutions to regression problems where the selected predictors appear in a structured pattern, with respect to a predefined distance measure over the candidate predictors. Our method is motivated by the problem of predicting HIV-1 drug resistance from protein sequences. We find that our method improves the interpretability of drug resistance while producing comparable predictive accuracy to standard methods. We also demonstrate our method in a simulation study and present some theoretical results and connections.