Exploring Saccharomycotina Yeast Ecology Through an Ecological Ontology Framework

Abstract

Yeasts in the subphylum Saccharomycotina are found across the globe in disparate ecosystems. A major aim of yeast research is to understand the diversity and evolution of ecological traits, such as carbon metabolic breadth, insect association, and cactophily. This includes studying aspects of ecological traits like genetic architecture or association with other phenotypic traits. Genomic resources in the Saccharomycotina have grown rapidly. Ecological data, however, are still limited for many species, especially those only known from species descriptions where usually only a limited number of strains are studied. Moreover, ecological information is recorded in natural language format limiting high throughput computational analysis. To address these limitations, we developed an ontological framework for the analysis of yeast ecology. A total of 1,088 yeast strains were added to the Ontology of Yeast Environments (OYE) and analyzed in a machine-learning framework to connect genotype to ecology. This framework is flexible and can be extended to additional isolates, species, or environmental sequencing data. Widespread adoption of OYE would greatly aid the study of macroecology in the Saccharomycotina subphylum.

Keywords: controlled vocabulary; dynamic; formal; isolation environment; macroecology; statistical enrichment.

FIGURE 1 |

Ontology subset describing the isolation environment of Metschnikowia mauinuiana. Each box represents a distinct class in the ontology. Each class is a subclass of a single class higher-up in the ontology. There are two relational properties shown in the figure (green and yellow arrows) that describe relationships between classes. The strain of M. mauinuiana shown is an instance (red arrow) of the specific environment from which it was isolated.

FIGURE 2 |

Relative distribution of the isolation environments in the ontology which includes 1088 yeasts. (A) The categories labeled “class” include yeasts that are an instance of that class or any of its subclasses. The categories labeled “modifier” are those connected to that class by a relationship. For example, any instance that contains the modifier “is from plant on animal” would be included in “Animal (modifier).” These classes are not exclusive—a yeast can be counted in both the “Plant” and “Angiosperm” categories. (B) Each order is divided into one of 5 exclusive categories, which are all classes. Therefore, no yeast is counted twice in this section. Not all yeasts, however, are classified into these groups. For example, there are 430 Serinales in this data set; due to the small overall number of samples, those sampled from other fungi are not shown.

FIGURE 3 |

The Ontology of Yeast Environments enabled machine learning analysis identify genes associated with specific environments. (A) The general framework for utilizing the yeast ecological ontology for machine learning. We identified a specific class of interested and obtained all the instances (yeast strains) either directly (black arrows) or relationally (colored arrows) associated with that class. The instances were then divided into training and testing datasets where the presence and absence of KEGG Orthologs (KOs) were used as features. We constructed a random forest and then interrogated the model for accuracy and the important features. (B) Classification of yeast in the animal class had an average AUC of 0.71 and an average true-positive rate of 66% across 100 iterations of the model. The KOs with the highest permutation importance are shown in the bar graph. (C) Classification of yeast in the plant class (including relational associated but with decayed plants removed) had an average AUC of 0.71 and an average true positive rate of 66% across 100 iterations of the model. There was a single KO (K09117) that had three times higher importance as the next most important KO.

FIGURE 4 |

KOs with known and unknown functions were highly informative in the construction of the random forest to classify yeast as isolated from plants or animals. The KOs associated with classification of yeasts in the animal or plant classes (first column) were clustered according to presence in the analyzed class (% Presence columns). The associated pathway for each KO is shown in column 2 with the two most important KOs (colored names) belonging to no known pathway. We also tested for statistical differences in the presence of the KOs in the yeasts belonging to the examined class as compared to those not in that class using a Fisher’s exact test. The p-value and odds ratio are reported in the last two columns and the raw data is presented in the FigShare repository.