The evolving species concepts used for yeasts: from phenotypes and genomes to speciation networks

Abstract

Here we review how evolving species concepts have been applied to understand yeast diversity. Initially, a phenotypic species concept was utilized taking into consideration morphological aspects of colonies and cells, and growth profiles. Later the biological species concept was added, which applied data from mating experiments. Biophysical measurements of DNA similarity between isolates were an early measure that became more broadly applied with the advent of sequencing technology, leading to a sequence-based species concept using comparisons of parts of the ribosomal DNA. At present phylogenetic species concepts that employ sequence data of rDNA and other genes are universally applied in fungal taxonomy, including yeasts, because various studies revealed a relatively good correlation between the biological species concept and sequence divergence. The application of genome information is becoming increasingly common, and we strongly recommend the use of complete, rather than draft genomes to improve our understanding of species and their genome and genetic dynamics. Complete genomes allow in-depth comparisons on the evolvability of genomes and, consequently, of the species to which they belong. Hybridization seems a relatively common phenomenon and has been observed in all major fungal lineages that contain yeasts. Note that hybrids may greatly differ in their post-hybridization development. Future in-depth studies, initially using some model species or complexes may shift the traditional species concept as isolated clusters of genetically compatible isolates to a cohesive speciation network in which such clusters are interconnected by genetic processes, such as hybridization.

Keywords: Comparative genomics; Fungi; Hybrids; Nomenclature; Species concepts; Taxonomy.

Fig. 1

Scheme showing increase of knowledge of yeast biodiversity with three levels of knowledge on certainty of species recognition

Fig. 2

Analysis of ITS1 + 2 and D1/D2 LSU ribosomal RNA gene sequences of yeasts. A, B Proportion of yeast ITS and LSU sequences at the class level. C, D Variation of the median ITS and LSU sequence similarity scores of yeasts at various taxonomic levels. E–H Optimal thresholds and associated highest F-measures predicted at the species and genus levels from a previous analysis (Vu et al. 2016) and a current dataset updated based on recent taxonomic revisions (date of analysis December 2020). The sequences were compared with each other using BLAST (Altschul et al. 1997). For each of the resulting local alignments of two sequences, a BLAST similarity score was calculated as the percentage of matches s if the alignment length l was greater than 300 bp (the minimum length of ITS sequences, Vu et al. 2019). Otherwise, the score was recomputed as $sl/300$. The names of the taxa associated with the sequences were downloaded from MycoBank (Robert et al. 2013)

Fig. 3

Taxonomic clustering of 4,356 fungal ITS barcodes of the current taxonomic classification. The sequences are coloured based on the genus name. Candida (with 593 sequences) is indicated in red, followed by Cryptococcus (329) in black, Saccharomyces (288) in blue, Rhodotorula (218) in green, Kluyveromyces (146) in cyan, Pichia (135) in pink, Debaryomyces (113) in yellow, Malassezia (106) in grey. The remaining 188 genera with less than 85 ITS sequences are coloured using the distinctColorPalette function of the randomcoloR package. The coordinates of the sequences were generated using fMLC (Vu et al. 2018). The sequences are visualised using the rgl R package (https://r-forge.r-project.org/projects/rgl/)