As an aside, it is interesting that unexpected nucleic acid binding from proteins is (yet again) a place where current structure prediction is falling short despite looking obvious in hindsight. There's probably many "dark matter" DNA/RNA binders that predictive tools are systematically blind to.