What's In John's Freezer?

Sorry about the title. It’s the best I could do. In case you missed it on our Anatomy to You blog, we unleashed a hefty database of CT (and some MRI) scans of our frozen crocodile cadavers last week, for free public usage. In total, it’s about 34 individuals from 5 species, in 53 databases constituting around 26,000 individual DICOM file format slices of data. This page has a table of what the data/specimens are. I am writing this post to share some more images and ensure that word gets out. We’re thrilled to be able to finally release this first dataset. We have plans to let loose a LOT more such data in the future, for various organisms that we study.

Stomach-Churning Rating: 2/10- be glad that these data don’t come with an olfactory component, especially the five rotten, maggot-ridden Morelet’s croc specimens, which are among the stinkiest things I’ve dealt with.

Crocodiles are no strangers to this blog, of course, as these past links testify. Indeed, most of the crocodile images I’ve blogged with come from specimens that are in this scan dataset. We even released a “celebrity crocodile, “WCROC” or FNC7 in our dataset, which is the 3.7m long Nile croc from “Inside Nature’s Giants”. It broke our CT scanner back in ~2009 but we got the data, except for the torso, and we also got some MRI scans from it, so we’re chuffed.

Above: The only spectacled caiman (Caiman crocodilus); and indeed the only alligatorid; in our dataset. To watch for: stomach contents/gastroliths, and all the damn osteoderms that I did/didn’t segment in this quickly processed file. This specimen had its limbs dissected for one of our studies, so only the right limbs are visible.

There are some more specimens to come- e.g. five baby Nile crocs‘ datasets (“GNC1-5”) are hiding somewhere in our drives and we just need to dig them up. You might also know that we published some scan data for crocodile vertebral columns (including fossils) in our recent paper with Julia Molnar et al. (and related biomechanical data discussed here), and we published all of our anatomical measurements for a huge set of crocodylian species in our papers by Vivian Allen et al. And then I had an enjoyable collaboration with Colleen Farmer and Emma Schachner on the lung anatomy of various crocodylian species, using these same specimens and related scan datasets.

Above: rotating Crocodylus moreletii (specimen FMC5 from our database) in a happy colour.

Sharing these kind of huge datasets isn’t so easy. Not only do few websites host them cheaply, and with reasonable file size limits, and limited headaches for what info you have to provide, and with some confidence that the websites/databases will still exist in 5-20 years, but also we were hesitant to release the dataset until we felt that it was nicely curated. Researchers can now visit my lab and study the skeletons (or in some cases, the still-frozen specimens) matched up with the scan data, and known body masses or other metadata. We’re not a museum with dedicated curatorial staff, so that was not trivial to reliably organize, and I still worry that somewhere in the dataset we mis-identified a specimen or something. But we’ve done our best, and I’m happy with that for now.

Above: rotating Osteolaemus tetraspis (specimen FDC2 from our database), which was obviously dissected a bit postmortem before we could scan it, but still shows some cool features like the extensive bony armour and the cute little doglike (to me, anyway) skull. I worked with these animals (live) a bit >10 years ago and came to love them. Compared to some other crocodiles we worked with, they had a pleasant demeanour. Like this guy:

Osteolaemus (resting) set up with motion capture markers for a yet-to-be-published gait study that we did in 2005 (ugh!). It wasn’t harmed by this.

Anyway, as a person who likes to maintain quality in the science we do, I also was hesitant to “just” release the DICOM file data rather than beautiful segmented 3D skeletal (or other tissue) geometry that is ready for 3D printing or animation or other uses, or interactive online tools like Sketchfab. Other labs (e.g. Witmerlab) do these kind of things better than we do and they inspire us to raise our game in the future, but I am sure that we will be forgiven for releasing big datasets without gorgeous visuals and more practical, processed files — this time. 🙂  We agree with many other scientists that sharing data is part of modern, responsible science– and it can be fun, too! Oddly enough, in this case we hadn’t used the CT/MRI data much for our own studies; most of the scans were never fully digitized. We just scan everything we get and figured it was time to share these scans.

Enjoy. If you do something cool with the data that we’ve made accessible, please let us know so we can spread the joy!

And if you’re a researcher headed to ICVM next week, I hope to see you there!