Pan-mutiomics uncovers the multidrug-resistant fungal determinants, enabling machine learning-based prediction of multidrug-hyperresistant strains from an outbreak fungal species complex

Strains from Cryptococcus gattii species complex (CGSC) caused the Pacific Northwest cryptococcosis outbreak, the largest cluster of life-threatening fungal infections in otherwise healthy human hosts known to date. Here, by using a pan-phenome-based method, we monitored the fitness outcomes of CGSC strains under 31 stress conditions, providing a large-scale overview of 2821 phenotype-strain associations in this important pathogenic clade. Phenotypic clustering analysis revealed a strong inter-correlation between distinct types of stress phenotypes in a proportion of CGSC strains, suggesting that shared determinants coordinate their adaptations to different stresses. In particular, a specific set of strains, including outbreak isolates, showed increased resistance to three antifungal drugs most commonly used to treat cryptococcosis. By integrating pan-genomic and pan-transcriptomic analyses, we predicted and determined the previously unrecognized genes that play pivotal roles in conferring multidrug resistance in an outbreak strain. From these genes, we found that expression variation of three genes as predictors was sufficient to predict multidrug hyper-resistant strains with AUC = 0.85, using machine learning algorithms. Overall, we developed a pan-mutiomics-based approach identifying cryptococcal multidrug-resistant determinants, which enables the prediction of hyper-resistant CGSC strains that may be of significant clinical concern.