Hongzhong Lu

This study uses yeast large-scale absolute quantitative proteomics to reveal universal principles of protein resource allocation among cellular organelles and its relationship to physiological parameters and species divergence.

scYeast is a foundation model for yeast single-cell transcriptomics that integrates biological priors, excelling in zero-shot tasks and transfer learning to other omics data.

AlphaGEM constructs genome-scale metabolic models using structural alignment and deep learning, improving accuracy over existing methods for diverse species.

MetaStrain combines enzyme-constrained models with meta-heuristic algorithms to efficiently identify combinatorial gene targets for improving microbial product yields.

StrainOptimizer improves microbial cell factory design by integrating multi-scale metabolic models with expression and proteome constraints, achieving higher bioproduction in yeast.

EnzymeTuning optimizes enzyme kinetic parameters in metabolic models using a GAN framework, improving proteome predictions and model performance across yeast species.

This study characterizes the genomic and metabolic adaptations of Saccharomyces cerevisiae across diverse niches using pan-genome and metabolic modeling.

Euktect improves eukaryotic sequence detection and classification in metagenomes using a deep-learning model, enhancing the identification of fungal species and linking eukaryotic sequences to clinical phenotypes.

This paper introduces a distributed large-scale neural network model for yeast transcriptional regulation, capable of predicting dynamic regulatory relationships and identifying key transcription factors.

DeepEnzyme improves kcat prediction using Transformer and GCN with sequence and 3D structure features, showing robustness for low similarity enzymes and mutation impact analysis.

Yeast9 improves yeast metabolic modeling by integrating multi-omics data, showing better prediction of growth rates and gene function than transcriptomics.

Standard-GEM provides a template for reproducible genome-scale metabolic model development following FAIR principles and software best practices.

The paper presents a computational pipeline to predict gene engineering targets for increased production of 102 chemicals in yeast using enzyme-constrained models.

A substrate graph neural network plus protein-sequence CNN predicts enzyme kcat (test Pearson r 0.71), tracks mutation effects, and parameterizes enzyme-constrained models for 343 yeast/fungi species.

FALCONET is an R package that automates the visualization of genome-scale metabolic models and multi-omics data using CellDesigner.