Recombination is a powerful evolutionary force that creates novel combinations from genetic machinery that has already been honed by natural selection. Although sexual reproduction is broadly ubiquitous among eukaryotes, many fungi have cryptic sexual cycles1. For instance, many fungal crop pathogens are not observed to reproduce sexually in nature, which has led to the assumption that adaptive genomic novelty arises predominantly from mutation and phenomena such as genomic rearrangements or horizontal gene transfer2. As a result, recombination is often neglected by the plant pathogen community3. Such has been the case for the fungus Pyricularia oryzae (synonymous with Magnaporthe oryzae; previously Magnaporthe grisea), a pathogen of cereal crops and both cultivated and wild grasses, for which field populations have previously been considered asexual4. Writing in Nature Ecology & Evolution, Rahnama et al.5 now challenge this expectation by using population genomics to show that recombination has been central to the rapid adaptation of P. oryzae to new hosts.

The potential for recombination to accelerate adaptation is perhaps greatest when it operates among divergent genetic groups — across either populations or species — and the idea that hybridization can enable the emergence of new crop diseases is well established6. Pyricularia oryzae is best known as the agent of rice blast, but since the late 1900s it has also caused wheat blast epidemics and, as such, is generally considered the most devastating fungal crop pathogen globally7. Rahnama et al. developed an elegant model that outlines how a series of admixture events between host-specialized P. oryzae individuals, culminating in a ‘multi-hybrid swarm’, enabled this host jump to wheat (Fig. 1). This was done using haplotype chromopainting: this method represents admixed (‘recipient’) haplotypes as a mosaic of colours, each of which is based on the single-nucleotide polymorphisms contained in divergent ‘donor’ haplotypes that were previously confined to one of several populations. The authors find that isolates sampled earlier (pre-swarm) have fewer contributors. Notably, they also find that negligible mutation has occurred in wheat-specialized individuals since the host jump, which suggests that recombination of existing genetic diversity was sufficient for ‘out of the box’ adaptation. Importantly, using Bayesian molecular clock analyses to date the divergence times of lineages, the authors report that this process probably happened within only 20 years — a notably short time frame.

Fig. 1: Evolutionary history of host jumps in the crop fungal pathogen P. oryzae.
figure 1

Rahmana et al. suggest that P. oryzae individuals from populations specialized to at least five wild-grass hosts (PoE, specialized to Eleusine, PoU to Urochloa, PoLu to Luziola, PoSt to Stenotaphrum and PoX to a mystery host lineage) were implicated in two hybridization events and then a multi-hybrid swarm, from which emerged P. oryzae individuals instantaneously specialized to rye (PoL, specialized to Lolium) and wheat (PoT, specialized to Triticum).

Full size image

Hybrid swarms (ongoing recombination among lineages) are delineated from one-off hybridization events8, but in both cases the resultant offspring (F1) are generally assumed to have reduced fitness in the parental niches. It is relatively straightforward to see the potential for further recombination to redistribute parental genetic diversity, but transgressive segregation also operates to break up co-adapted haplotypes and can produce hybrid (F2) phenotypes that are far more diverse than that of either parent (that is, transgressive segregants)9. These ‘hopeful monsters’ are extremely genetically diverse and favoured for rapid adaptation to new environmental conditions. As Rahnama et al. demonstrate, hybrid swarms may be particularly important for host jumps between crop wild relatives and crops6 — even across non-fungal pathogen lineages10. Low genetic diversity in a crop host combined with a novel pathogen that can then switch to clonal reproduction facilitates the emergence of a single, predominant pathogen genotype that would otherwise have disappeared if it were obligately sexual or if its host were outbred.

If recombination is both prevalent and powerful in fungal pathogens, the question arises of why we so frequently observe clonal field populations: probably because clonality is more efficient and less costly for expanding across a crop monoculture11. Nevertheless, the results of Rahnama et al. demonstrate that sexual reproduction does not need to happen frequently to have major implications, even speculating that the single swarm event could have occurred over mere weeks. The timing, and even infrequency, of punctuated sexual reproduction may actually be central to ensuring that recombination increases fitness, rather than breaking up already favourable combinations12. We may need to shift from assuming persistent clonality to assuming intermittent sexual reproduction13 — the questions might rather be ‘how much’ and ‘with whom’14.

How globalization can increase opportunities for fungi to hybridize is an ecological concern6. Phytosanitary regulations operate to minimize introductions of recognized pathogens, but they may not sufficiently regulate the movement of fungi that are not currently recognized as crop pathogens. Expanding pathogen ranges owing to global change factors will also increase coexistence and the potential for novel hybridization. New surveillance technology15 will be key to monitoring the occurrence of fungal populations and hopefully preventing or at least anticipating disease emergence. Surveillance of fungal pathogens across landscapes could also be used to identify wild-host reservoirs, and thus help to predict the likelihood of hybridization events and inform more-targeted crop-treatment measures.

The findings from Rahnama et al. also have implications for how we engineer crop resistance. For both tractability and repeatability, conventional resistance gene discovery has undoubtedly benefitted from clonal pathogen lineages to detect resistance genes. However, the strength of using clonal pathogens to identify resistance has perhaps blinded us to the mechanisms by which pathogens evolve and adapt to overcome host resistance in nature. Although omitting recombination makes resistance genes easier to identify, those resistance genes will be tested in an environment in which it may be operating. Even stacked resistance genes can be circumvented by recombination, especially amongst high levels of genetic diversity within pathogen populations16. Indeed, Rahnama et al. provide evidence that it was the inheritance of non-functional alleles for genes that determine avirulence to wheat that enabled the host jump in P. oryzae. Although the authors emphasize the contributions of a hybrid swarm to this process, it is worth remembering that even individual recombination events among closely related fungi can add or remove virulence genes and thus facilitate invasion of a novel host or variety. It is informative to look for signals of hybrid swarms to develop our understanding of crop–pathogen adaptation, but these processes may also be operating at a much more subtle level to drive novel host colonization. Knowledge of these mechanisms could help to inform durable resistance deployment strategies (for example, rotation, pyramiding or stacking, mixture, and mosaic)3,17.

Rahmana et al. neatly demonstrate the considerable impact of recombination in novel disease emergence in a major fungal pathogen of one of the most widely cultivated crops. However, these processes are undoubtedly far more widespread than is currently appreciated. A better understanding of the contributions of recombination across diverse crop–pathogen systems are needed, especially in the context of co-occurring crop wild relatives. A number of fungal species with population-level genomic data represent ideal systems in which to further explore the differential effects of mutation and recombination on the rate of adaptation to new hosts and environments18. With these improved models of how fungal crop pathogens evolve in nature, we can hope to move from explaining past disease emergence towards preventing future epidemics.