Aspergillus fumigatus is one of the most common pathogenic fungi, which results in high morbidity and mortality in immunocompromised patients. Amphotericin B (AMB) is used as the core drug for the treatment of triazole-resistant A. fumigatus. Following the usage of amphotericin B drugs, the number of amphotericin B-resistant A. fumigatus isolates showed an increasing trend over the years, but the mechanism and mutations associated with amphotericin B sensitivity are not fully understood. In this study, we performed a k-mer-based genome-wide association study (GWAS) in 98 A. fumigatus isolates from public databases. Associations identified with k-mers not only recapitulate those with SNPs but also discover new associations with insertion/deletion (indel). Compared to SNP sites, the indel showed a stronger association with amphotericin B resistance, and a significant correlated indel is present in the exon region of AFUA_7G05160, encoding a fumarylacetoacetate hydrolase (FAH) family protein. Enrichment analysis revealed sphingolipid synthesis and transmembrane transport may be related to the resistance of A. fumigatus to amphotericin B. The expansion of variant types detected by the k-mer method increases opportunities to identify and exploit complex genetic variants that drive amphotericin B resistance, and these candidate variants help accelerate the selection of prospective gene markers for amphotericin B resistance screening in A. fumigatus.

The cultivation of hybrids with favorable complex traits is one of the important goals for animal, plant, and microbial breeding practices. A method that can closely predict the production performance of hybrids is of great significance for research and practice. In our study, polygenic risk scores (PRSs) were introduced to estimate the production performance of Saccharomyces cerevisiae. The genetic variation of 971 published isolates and their growth ratios under 35 medium conditions were analyzed by genome-wide association analysis, and the precise p-value threshold for each phenotype was calculated. Risk markers for the above 35 phenotypes were obtained. By estimating the genotype of F1 hybrids according to that of the parents, the PRS of 613 F1 hybrids was predicted. There was a significant linear correlation between the maximum growth rate at 40 °C and PRS in F1 hybrids and their parents (R2 = 0.2582, R2 = 0.2414, respectively), which indicates that PRS can be used to estimate the production performance of individuals and their hybrids. Our method can provide a reference for strain selection and F1 prediction in cross-breeding yeasts, reduce workload, and improve work efficiency.