Introduction
The genus Saccharomyces (Saccharomycetales, Ascomycota) originated ~100 million years ago facilitated by whole genome duplication and an increased metabolic capacity to degrade sugar to ethanol . It currently comprises eight species, which diverged 4-20M years ago . All species are monophyletic with high levels of sequence collinearity but show vast genetic and ecological diversity . Recently, the genomic and ecological data available for all eight species (and their hybrids) have considerably grown , promoting yeast from a traditional laboratory model system that only included a few clonal strains to one in which we can answer questions relevant for ecology and evolution . For instance, we can now resolve the genetic basis of adaptation to environmental challenges , we can investigate which ecological factors determine the distribution and structure of yeast populations in the wild , and measure the level of divergence and gene flow between them . Our growing knowledge of the phenotypic and genetic diversity of wild yeasts also allows us to study important ecological traits comparatively, across species backgrounds, alleviating limitations resulting from only including a handful of well-characterized laboratory strains in our analyses. One ecological trait that stands out as particularly diverse between species are their temperature preference profiles . Resistance and adaptation to high or fluctuating temperatures are an increasing research focus for climate change biology. Temperature profiles are also important for the development of industrial strains, e.g. for fermentation products and biofuels . Of the eight species included here, five are considered cold-tolerant (S. kudriavzevii, S. arboricola, S. uvarum , S. eubayanus , and the recently discovered S. jurei) , one species is thermo-tolerant (S. cerevisiae ), and two species are considered a thermo-generalist (S. paradoxus andS. mikatae ) growing well in a broad range of temperatures .
Developing wild strains into effective systems for research and industry requires the systematic testing and measuring of fundamental population phenotypes including their growth rates, kinetics (e.g. the length of the lag phase), and yield. Microbial research often applies high-throughput methods to estimate population growth and fitness in environments of interest, e.g. media containing different nutritional and stress conditions. A common technique is to measure the optical density (OD) of microbial cultures using a spectrophotometer (‘plate reader’). Optical density measures the turbidity of liquid cultures, which is assumed to be proportional to the cell number, i.e. the concentration of cells in the sample . Specifically, OD is the negative log of transmittance, i.e. the fraction of light detected when passed through a cuvette or micro-titer plate containing the microbial culture. It is typically measured at a wavelength of 600nm as this electromagnetic radiation is thought to not cause cell damage. Calculations follow the Beer-Lambert law , which states that OD is proportional to the concentration of a solution. However, this law only applies to cultures with low cell densities (typically OD600 up to 0.1). At higher densities, the light gets increasingly scattered between cells, and OD does not increase as fast as the cell titer. Using spectrophotometry to infer population fitness has additional limitations affecting the translation of OD into cell counts. Importantly, the method does not differentiate between dead and alive cells and the absorption coefficient (ε) can be affected by cell size (Fukuda 2023). Different methods exist for the calibration of OD measurements , including the use of silica microspheres, direct cell counting with microscopy, and colony counting in serial dilutions on agar plates . But the most efficient, high-throughput method is flow cytometry, using laser-based detection of individual cells that allows for accurate cell count estimates
Here, we explore the variation in the relationship between OD measures from spectrophotometry and cell counts from flow cytometry, across all eight Saccharomyces yeast species. To also test for variation within species, we used three strains per species (except for S. jurei , where only two strains were available) from different geographic and ecological origins, including isolates from fruit, soil, rotten wood, and tree bark from Europe, Asia, North and South America, and Australia (Table S1). We hope our study expands the knowledge base of important growth parameters of non-cerevisiae strains and improves the biological interpretation of population fitness data from wild, non-domesticated yeasts.