Song‐Min Schinn

LeGenD predicts N-glycan structures and their abundance using lectin-binding patterns and AI, trained on glycoprofiles from recombinant proteins.

GlyCompareCT preprocesses glycomics data to address sparsity and non-independence, improving statistical power for downstream analysis.

The paper evaluates parameters for genome-scale metabolic models of cultured mammalian cells, finding that homeostasis, dynamic morphology, and biomass composition are critical for accurate predictions.

The paper integrates machine learning with genome-scale metabolic models to accurately predict amino acid concentrations in Chinese Hamster Ovary cell cultures over time.

The paper identifies protein-protein interactions supporting recombinant protein secretion using BioID and Bayesian modeling, finding key roles for oxidative folding enzymes in disulfide bond formation.

StanDep improves context-specific metabolic models by using transcriptomic variability to identify core reactions, increasing inclusion of housekeeping and lowly expressed reactions.