Final Report Summary - GEDINDIS (Genomics of Ecological Divergence in the fungal species Neurospora discreta) The main motivation of my research is to understand the processes driving the transition from local adaptation to incipient species, which I call the continuum of adaptive divergence. I use the distribution of genetic divergence within the genomes of differentiating lineages as a key observable feature of this phenomenon. I focus my research on fungi because they have small and dense genomes and they are amenable to experimental approaches. They also provide ideal systems to test of the universality of dogma derived from studies of other eukaryotes.The difficulty in working with fungi has been identifying adaptive phenotypes beyond simple metabolic traits. High-throughput sequencing techniques make it possible to obtain a large portion of the genome of any individual of any species at an affordable cost. This allows for «reverse adaptive genetics» that is, to identify genes with unusual evolution, to identify environmental parameters consistent with the function of the genes, and then test hypotheses by genetic manipulation.The study of Ellison et al. (2001, PNAS 108:2831) provides an elegant application of this approach on the model eukaryote Neurospora - a filamentous fungus that can be found growing on dead plant matter after fires, in tropical, subtropical and temperate regions. Ellison et al. scanned for divergence between two latitudinally distinct populations of the species N. crassa (Caribbean basin and Louisiana). They found two highly divergent genomic islands differentiating the two populations, with genes inside distinct genomic islands having functions and patterns of variation consistent with local adaptation to day length and temperature. Growth-rate assays suggested the divergence islands may be the result of local adaptation to the 9 °C difference in average yearly minimum temperature between these two populations.To challenge the hypothesis developed for speciation in N. crassa I proposed to study the origin of ecological differentiation in a second species with an even greater latitudinal distribution, N. discreta phylogenetic species 4 subgroup B (N. discreta PS4B). The project objectives were: 1) to infer the demographic history of N. discreta PS4B, 2) to scan for adaptive divergence between differentiated populations of N. discreta PS4B. Data collection.I chose 52 samples from the N. discreta PS4B culture collection available in the host lab, representing 7 locations in America, Europe and Asia for N. discreta PS4B [Thailand, New Mexico (NM), California (CA), Washington (WA), Alaska (AK), Spain and Switzerland (EU)] and a single site in Papua New Guinea for the sister group N. discreta PS4A. I also used an isolate of N. discreta PS6 from Côte d’Ivoire as an outgroup. Illumina sequencing libraries were prepared using the Illumina sample preparation kit, from total DNA extracted from the mycelia and purified using Qiagen kits. High-throughput Illumina sequencing was performed using the HiSeq 2500 sequencer of Berkeley sequencing facility. I built an analysis pipeline making use of freely available programs (Prinseq, Bowtie2, GATK) to map reads on the genome sequence of N. discreta PS4B and call SNPs. The final dataset included ca. 1.2 Million high-quality SNPs of which ca. 380,000 were observed in N. discreta PS4B.Data analysis.I analyzed population subdivision based on the full set of SNPs using a combination of Maximum likelihood phylogenetic inference and model-based clustering methods. The N. discreta PS6 isolate was used an outgroup, and I found 2 basal lineages in Thailand and Papua New Guinea, and 3 more derived lineages in North-America/Europe. Individuals from AK and EU on the one hand, and from NM on the other hand grouped into sister lineages, and individuals from CA were basal to these two lineages. Individuals from WA, sampled within the same site, grouped with either the NM individuals, or the CA individuals, indicating the coexistence in sympatry of two divergent lineages. Lineages NM-WA and CA-WA may have recolonized North-America from glacial refugia during the current interglacial. The AK-EU lineage may result from human-mediated transportation or may correspond to a circumpolar species, as it is known to exist in some plant-associated fungi, like lichens. No hybrid was found in WA, were two divergent lineages occupying the same habitat (the surface of burned vegetation) co-occur in sympatry. However, one individual from AK had ca. 10% ancestry in the NM-WA lineage, providing the first direct evidence for the existence of Neurospora hybrids in nature.Having found evidence for contemporary hybridization, we tested for historical hybridization using a method devised by Green et al. (2010 Science 328: 710), which compares two classes of shared derived alleles, termed ABBAs and BABAs. Estimated values of the D statistic, which tests for a significant imbalance of ABBAs and BABAs, was indicative of past hybridization between NM-WA and CA-WA lineages. However, an alternative explanation that cannot be ruled out is ancient subdivision of the ancestral population. We also used a likelihood model-based approach for demographic inference using diffusion approximation to the allele frequency spectra of diverging populations (dadi program). Analyses revealed that a ((NM-WA,AK-EU),CA-WA) model of strict isolation and introgression following secondary contact was the most likely, with varying rates of introgression generating genomic heterogeneity in levels of divergence between lineages. Parameter estimates indicated that the different lineages diverged at approximately the same time (i.e. the model is almost a trifurcation), ca. 440,000 generations ago. Inferred rates of gene flow were low, with a median of effective migration rates on the order of 1e-1.Population genetic analyses were thus consistent with gene-flow among lineages, although recent and very limited in magnitude. We therefore decided to carry out in vitro experiments to determine if the different lineages are able to cross in lab conditions. These experiments did not reveal strong reproductive isolation between the different lineages and suggest they are isolated by one of several of the following mechanisms: temporal isolation (i.e. different mating times), habitat isolation (i.e. different mating sites), ecological barriers (i.e. immigrant inviability, selection against hybrids), intrinsic late postzygotic barriers (i.e. hybrid sterility), or another form of intrinsic early postzygotic isolation (i.e. hybrid inviability).Unlike the study of Ellison et al. (2001, PNAS 108:2831), genome scans for divergence and genome scans for natural selection did not reveal genomic features that could be resistant to gene flow.Population genomics offers great potential for enhanced understanding of the population biology and evolution of fungi. Many putative gene transfer events between phylogenetically divergent fungal lineages have been discovered, and our work highlights the quantitative importance of genetic exchanges between more closely related taxa to the evolution of fungal genomes. Our study also supports the role of allopatric isolation as a driver of diversification in saprobic microbes. Similar investigations on other models will allow quantify how common are the factors uncovered in our study in the diversification of fungal lineages. My findings challenge the universality of dogmatic views derived from studies of animals and plants regarding the evolutionary significance of interspecific gene flow. Advances in our understanding of fungal genomics is also relevant for global change, emergence and spread of invasive species and pathogens, agricultural and forestry production.
