Comparative genomic analysis of clinical Candida glabrata isolates identifies multiple polymorphic loci that can improve existing multilocus sequence typing strategy

Candida glabrata is the second leading cause of candidemia in many countries and is one of the most concerning yeast species of nosocomial importance due to its increasing rate of antifungal drug resistance and emerging multidrug-resistant isolates. Application of multilocus sequence typing (MLST) to clinical C. glabrata isolates revealed an association of certain sequence types (STs) with drug resistance and mortality. The current C. glabrata MLST scheme is based on single nucleotide polymorphisms (SNPs) at six loci and is therefore relatively laborious and costly. Furthermore, only a few high-quality C. glabrata reference genomes are available, limiting rapid analysis of clinical isolates by whole genome sequencing. In this study we provide long-read based assemblies for seven additional clinical strains belonging to three different STs and use this information to simplify the C. glabrata MLST scheme. Specifically, a comparison of these genomes identified highly polymorphic loci (HPL) defined by frequent insertions and deletions (indels), two of which proved to be highly resolutive for ST. When challenged with 53 additional isolates, a combination of TRP1 (a component of the current MLST scheme) with either of the two HPL fully recapitulated ST identification. Therefore, our comparative genomic analysis identified a new typing approach combining SNPs and indels and based on only two loci, thus significantly simplifying ST identification in C. glabrata. Because typing tools are instrumental in addressing numerous clinical and biological questions, our new MLST scheme can be used for high throughput typing of C. glabrata in clinical and research settings.