Soha Hassoun

Fine-Grained Structural Classification of Biosynthetic Gene Cluster-Encoded Products

The paper introduces BGCat, a machine learning tool for fine-grained classification of biosynthetic gene cluster-encoded products, improving upon existing methods and labeling over 100k BGCs.

Learning from All Views: A Multiview Contrastive Framework for Metabolite Annotation

This paper introduces MVP, a multiview contrastive framework for metabolite annotation that improves molecular candidate ranking using multiple data views.

FLARE: Fine-grained Learning for Alignment of spectra-molecule REpresentation Enhances Metabolite Annotation

FLARE improves metabolite annotation by aligning spectra and molecular structures through contrastive learning, achieving state-of-the-art performance on MassSpecGym.

Protein-ligand co-design: a case for improving binding affinity between Type II NADH:quinone oxidoreductase and quinones

The paper introduces a co-design paradigm to enhance binding affinity between Ndh2 and quinones, demonstrating its application in improving biosensing capabilities.

Molecular structure discovery for untargeted metabolomics using biotransformation rules and global molecular networking

This paper introduces BAM, a method using biotransformation rules to annotate untargeted metabolomics spectra, achieving 24.2% correct structure assignments in a large molecular network.

Computational Analysis of the Gut Microbiota-Mediated Drug Metabolism

This study introduces a computational framework, MDM, to predict gut microbiota-mediated drug metabolism by integrating diverse datasets and existing tools to identify potential metabolites and their associated enzymes.

Evaluation of search-enabled Pre-trained Large Language Models on retrieval tasks for the PubChem Database

This study evaluates GPT-4o's ability to retrieve data from PubChem using natural language and programmatic access, showing programmatic retrieval is more accurate.

Analysis of Metabolic Network Disruption in Engineered Microbial Hosts due to Enzyme Promiscuity

The paper develops a computational method to identify enzyme promiscuity-induced metabolic network disruptions in engineered microbial hosts and evaluates its application in two case studies.

Towards creating an extended metabolic model (EMM) for E. coli using enzyme promiscuity prediction and metabolomics data

The paper predicts new metabolites and reactions in E. coli using enzyme promiscuity and metabolomics data, enhancing the iML1515 metabolic model.

Establishing Synthesis Pathway-Host Compatibility via Enzyme Solubility

This paper introduces ProPASS, a workflow that links pathway synthesis with enzyme-host compatibility using solubility scores, and presents ProSol DB, a database of solubility scores for 240,016 enzymes in E. coli.