H. Nakazawa

This study uses low-flux electron diffraction to analyze intercellular lipid organization in human lip stratum corneum, finding higher fluid-phase lipid proportions linked to lower water retention and chapping vulnerability.

This study shows that ML-guided directed evolution can lead to different enzyme variants depending on the presence of a highly positive variant in training data, achieving similar activity improvements.

The paper introduces a machine-learning-guided mutagenesis platform that accelerates discovery of functional proteins by using two rounds of mutagenesis informed by a trained model.