Guomin Han

The paper introduces an annotation-independent machine learning framework to identify functional genes across species for shared phenotypes, demonstrating effectiveness in plant-mycorrhizal symbiosis, vertebrate skeletal development, and bacterial traits.

The paper uses machine learning to predict bacterial shape from genomic data and identifies key genes involved in rod-shaped morphology through experimental validation.

The paper develops a method to extract endophytic microbiome information from plant and animal transcriptome data, providing microbial composition alongside host gene expression.