Tessellation

Man, it was a fun day it the lab yesterday. Kharis, Sarah, and I spent the whole morning working on a project together, and we managed to make great progress. 

Recently, Kharis has been struggling to identify larvae from the Arctic. Don't get me wrong - we have a good methodology, thanks to a few years of struggling and about 6 months of Johanna troubleshooting. We can reliably get good-quality sequences from our specimens now. The DNA was not the problem (for once).

Kharis's struggle has been morphology. She sorted all her larvae into categories based on how they looked in the field. It's a strategy I taught her - sort everything live, photograph and preserve the specimens, then sequence them back in the lab. By sorting up-front, you can save yourself a lot of time and just sequence a few representatives of each morphotype once you're home. If the sequences for those representatives line up, you have a reliable identification for the morphotype. 

If. 

See, that's the problem. Kharis's "Pluteus 1" came back as two different species. "Trochophore clear" was at least three. Animals that looked identical in the field were turning out to be different species. It was super frustrating, and we were afraid we would have to end up sequencing every single specimen. 

Before throwing in the towel and spending a ton of money on sequencing, I suggested we try one more thing. Kharis photographed each larva in the field, remember. What if specimens that looked identical to her tired, overwhelmed field work brain were actually subtly different? Maybe if we printed out all the photos now, we could tell the species apart. 

We printed the photos. We labeled them individually. We cut them apart. We taped them to the white board. We stood back to get a good look. And then we sorted them. The lab became one giant tessellation with rectangular photos in orderly stacks and rows. 

By the end of the day, "Pluteus 1" had become 3 different morphotypes, and "Trochophore clear" was 6 or 8. We'll have to sequence more morphotypes now, but at least we can trust that all individuals in a morphotype are the same. We have confidence in our data again. 

Cheers for scientific progress!

Sarah and Kharis showing off their sorted photos.
Every pile on the floor represents a morphotype. 
Images on the white board are sorted "pluteus 1."