Zisis Kozlakidis

This paper benchmarks AI-based drug-resistance prediction models for Mycobacterium tuberculosis using a large unified dataset, finding XGBoost performs best in precision-recall and F1 metrics, while neural models excel in AUROC.

This study identifies HVR-1 as an autoinhibitory region in HCV E1E2 that suppresses viral entry through conformational entropy, which is released upon receptor binding to enhance entry but reduce antibody resistance.

The paper analyzes HIV-1 full-genome sequences from London to reveal high strain variability, including unique recombinants and transmission clusters not detectable with partial sequencing.