J. Otwinowski

Learning the shape of protein micro-environments with a holographic convolutional neural network

This paper introduces H-CNN, a machine learning model that predicts mutation impacts on protein function using structural and evolutionary data.

Fierce selection and interference in B-cell repertoire response to chronic HIV-1

The paper analyzes B-cell repertoire evolution in chronic HIV-1 infection, finding that beneficial mutations are frequently outcompeted by new mutants due to clonal interference.

Inferring the shape of global epistasis

The paper infers the shape of global epistasis in three proteins using high-throughput mutagenesis data and a maximum-likelihood method with monotonic nonlinear functions.