Deep learning image analysis for filamentous fungi taxonomic classification: Dealing with small datasets with class imbalance and hierarchical grouping.

Deep learning applications in taxonomic classification for animals and plants from images have become popular, while those for microorganisms are still lagging behind. Our study investigated the potential of deep learning for the taxonomic classification of hundreds of filamentous fungi from colony images, which is typically a task that requires specialized knowledge. We isolated soil fungi, annotated their taxonomy using standard molecular barcode techniques, and took images of the fungal colonies grown in petri dishes (n = 606). We applied a convolutional neural network with multiple training approaches and model architectures to deal with some common issues in ecological datasets: small amounts of data, class imbalance, and hierarchically structured grouping. Model performance was overall low, mainly due to the relatively small dataset, class imbalance, and the high morphological plasticity exhibited by fungal colonies. However, our approach indicates that morphological features like color, patchiness, and colony extension rate could be used for the recognition of fungal colonies at higher taxonomic ranks (i.e. phylum, class, and order). Model explanation implies that image recognition characters appear at different positions within the colony (e.g. outer or inner hyphae) depending on the taxonomic resolution. Our study suggests the potential of deep learning applications for a better understanding of the taxonomy and ecology of filamentous fungi amenable to axenic culturing. Meanwhile, our study also highlights some technical challenges in deep learning image analysis in ecology, highlighting that the domain of applicability of these methods needs to be carefully considered.

Surfactant-Mediated Effects on Hydrological and Physical Soil Properties: Data Synthesis.

Soils are under the threat of a multitude of anthropogenic factors affecting the complex interplay of various physical and hydrological soil processes and properties. One such factor is the group of surface-active compounds. Surfactants have a broad range of applications and can reduce solid-liquid interfacial forces and increase wettability and dispersion of particles. Surfactant effects are context-dependent, giving rise to a wide range of reported effects on different soil processes and properties. Here, we evaluate the evidence base of surfactant research on 11 hydrological and physical soil variables. Our goal was to identify knowledge gaps and test the robustness of the proposed surfactant effects. We found that the current knowledge base is insufficient to reach strong data-backed conclusions about the effects of surfactants in soils. We identified a unique case of bias in the data as a result of conflated patterns from laboratory and field studies. We could not support the hypothesis that the surfactant charge determines soil effects for any of the tested soil variables. We believe that further experiments on surfactant-mediated effects on soil properties and processes are urgently required, paying attention, in particular, to improving experimental design and data reporting standards.

Root associated fungal lineages of a tropical montane forest show contrasting sensitivities to the long-term addition of nitrogen and phosphorus.

Root associated fungal (RAF) communities can exert strong effects on plant communities and are potentially sensitive to shifts in soil fertility. As increased atmospheric nitrogen (N) and phosphorus (P) deposition can alter the nutrient balance in natural ecosystems, we assessed the response of RAF communities to a fertilization experiment deployed on a highly diverse Andean forest. The stand level fine root fraction was sampled after 7 years of systematic N and P additions and RAF communities were characterized by a deep sequencing approach. We expected that fertilization will enhance competition of fungal taxa for limiting nutrients, thus eliciting diversity reductions and alterations in the structure of RAF communities. Fertilization treatments did not reduce RAF richness but affected community composition. At the phylum level fertilization reduced richness exclusively among Glomeromycota. In contrast, N and P additions (alone or in combination) altered the composition of several fungal phyla. The lack of a generalized response to long-term fertilization among RAF lineages suggests that most of these lineages will not be directly and immediately affected by the increasing rates of atmospheric N and P deposition expected for this region by 2050.

Moderate phosphorus additions consistently affect community composition of arbuscular mycorrhizal fungi in tropical montane forests in southern Ecuador.

Anthropogenic atmospheric deposition can increase nutrient supply in the most remote ecosystems, potentially affecting soil biodiversity. Arbuscular mycorrhizal fungal (AMF) communities rapidly respond to simulated soil eutrophication in tropical forests. Yet the limited spatio-temporal extent of such manipulations, together with the often unrealistically high fertilization rates employed, impedes generalization of such responses. We sequenced mixed root AMF communities within a seven year-long fully factorial nitrogen (N) and phosphorus (P) addition experiment, replicated at three tropical montane forests in southern Ecuador with differing environmental characteristics. We hypothesized: strong shifts in community composition and species richness after long-term fertilization, site- and clade-specific responses to N vs P additions depending on local soil fertility and clade life history traits respectively. Fertilization consistently shifted AMF community composition across sites, but only reduced richness of Glomeraceae. Compositional changes were mainly driven by increases in P supply while richness reductions were observed only after combined N and P additions. We conclude that moderate increases of N and P exert a mild but consistent effect on tropical AMF communities. To predict the consequences of these shifts, current results need to be supplemented with experiments that characterize local species-specific AMF functionality.

The relative importance of ecological drivers of arbuscular mycorrhizal fungal distribution varies with taxon phylogenetic resolution.

The phylogenetic depth at which arbuscular mycorrhizal (AM) fungi harbor a coherent ecological niche is unknown, which has consequences for operational taxonomic unit (OTU) delineation from sequence data and the study of their biogeography. We tested how changes in AM fungi community composition across habitats (beta diversity) vary with OTU phylogenetic resolution. We inferred exact sequence variants (ESVs) to resolve phylotypes at resolutions finer than provided by traditional sequence clustering and analyzed beta diversity profiles up to order-level sequence clusters. At the ESV level, we detected the environmental predictors revealed with traditional OTUs or at higher genetic distances. However, the correlation between environmental predictors and community turnover steeply increased at a genetic distance of c. 0.03 substitutions per site. Furthermore, we observed a turnover of either closely or distantly related taxa (respectively at or above 0.03 substitutions per site) along different environmental gradients. This study suggests that different axes of AM fungal ecological niche are conserved at different phylogenetic depths. Delineating AM fungal phylotypes using DNA sequences should screen different phylogenetic resolutions to better elucidate the factors that shape communities and predict the fate of AM symbioses in a changing environment.