Caveats and Future Directions
Consistent with the prediction that more rainfall results in more pathogens, which result in more disease resistance – we find a positive correlation between R gene counts and precipitation. However, the distance between our observed correlation and our causal motivation is quite large. Here we suggest two routes to bridge this gap – independent replication and unraveling the causal chain.
Correlation does not imply causation. However, repeated independent replication “does waggle its eyebrows suggestively and gesture furtively while mouthing ’look over there’.” (Munroe 2009). Thus, a key step in establishing that the correlation uncovered reflects our biological hypothesis, rather than happenstance, would be to evaluate the generality of this pattern. Some such evidence already exists – similar results have been seen in RFLP based studies of a few R genes in big bluestem and switchgrass (Rouse et al 2011, Zhu et al 2013). Testing for this pattern in other Silphium species represents a promising direction, as they would provide evolutionarily independent replication, while covering a similar precipitation gradient, and could use the same RenSeq baits developed here. Extending this study to more distant taxa would provide further evidence supporting this hypothesis.
Functional studies of these R-genes would provide more evidence for our motivating causal hypothesis. A complete, phased, chromosome-level assembly of S. integrifolium will both allow for better assessment of whether these genes are functional, and enable association studies to determine the loci of pathogen resistance for incorporation into breeding programs. Additional evaluation of the hypothesis that the number of pathogens affecting Silphium (and/or the variation in their ability to evade a specific R gene), as well as associating specific NLR alleles to resistance to specific pathogens would allow for a more mechanistic understanding of the association uncovered in this paper.