Jeffrey W. Miller

This paper shows that LOESS and DE-SWAN can create false 'waves' of molecular aging in data, challenging their statistical validity.

This paper introduces a Bayesian NMF approach with power posteriors and sparsity-inducing priors to improve mutational signature discovery, showing better accuracy than existing methods.

The paper introduces scGBM, a method for dimensionality reduction in scRNA-seq data using a Poisson bilinear model, which better captures biological information and quantifies uncertainty.

The paper compares classification algorithms to predict pro- and anti-longevity genes using gene expression and ontology data in C. elegans and S. cerevisiae, finding elastic net logistic regression performs best.