One highlight of #SSB2015 was a series of debates organized to focus on key points of uncertainty or debate in the field of systematics and phylogeny: Better models vs. more data (Hillis vs. Rokas), species concepts (Simon vs. Mallet), and molecular divergence times (Marshall vs. Hedges). Cécile Ané and I were asked to debate comparative methods, but a debate did not suit us, and so we built a presentation together about the limits of comparative methods.
A successful but immature field shouts “Wow, look at what we can do!” A mature field ponders “What are the limits of what we can know?” After decades of almost unchecked enthusiasm, fuelled by oceans of new data and computational capabilities, we are waking up to our limits. The success in reconstructing phylogeny has perhaps made us overconfident, believing that we should have similar success in phylogenetic studies of ancestral states, character correlation and diversification. However, the optimism shouldn’t transfer: reconstructions of phylogeny itself gain their power from the entire genome, while methods using phylogeny to answer evolutionary questions usually have sample sizes limited to the number of species at best.
The presentation Cécile and I put together asked: What are our humbling limits in principle, in the style of Heisenberg’s Uncertainty and Gödel’s Undecidability? What are our unfortunate limits in practice, given the amount of data we can usually gather? I’ll try to post a few more snippets from it.