Metabarcoding is incredibly powerful for diet analysis—but it has a well-known Achilles’ heel: host DNA swamping.
When predator and prey are genetically similar (e.g., feral pigs consuming other vertebrates), samples can come back ~100% host. At that point, you can’t tell whether vertebrate prey is absent—or just completely masked.
We tried “blocking” primers. They reduce host amplification, but rarely enough to recover meaningful prey signal. Increasing sequencing depth (amplicon or shotgun) didn’t solve it either. Even switching primers to avoid host (or gut bacteria) could reveal some prey—but we could always be missing other prey when the potential prey set is broad.
This issue seemed intractable for years, but if there’s one thing you learn in Boulder, there’s more than one way up a mountain.
So we asked: what would happen if we had a technique where we could add RNA complementary to target sequences, allow hybridization, wash away all the other DNA (like the host DNA), and sequence what remains.
This approach is essentially targeted capture with RNA baits. It’s a technique used frequently in health studies, evolutionary studies, and a few ancient DNA studies, but never really for wildlife diet studies.
For feral pig diet, targeted capture worked remarkably well. Using a ~20k bait set spanning vertebrates, we went from seeing only host DNA to recovering full-length COI sequences from prey and reducing host sequencing to just 2%. We never would have known feral pigs were eating pygmy rattlesnakes otherwise.
More recently, we designed a bony fish bait set targeting 12S and COI and applied it to shark diet samples (sandbar shark (Carcharhinus plumbeus), shown below).
What we saw:
~1000-fold enrichment of targets
Recovery of full-length COI and 12S markers
Some enrichment of host, but not enough to obscure signals.*
*refining the bait set should help here in the future
With targeted capture, there is always some off-target carry-over, which you can see in the ~5x coverage of host and prey, which functions like genome skimming and let’s us reconstruct full mitogenomes for host and prey.
We still have some refinements to make, but targeted capture is fundamentally changing how we approach diet reconstruction at Jonah.
We can now design bait sets for any informative marker across taxa and recover full-length sequences, dramatically improving taxonomic resolution.
With targeted capture, prey is no longer hidden behind host.