Nrecon 1

For general tooth morphology, heads were fixed overnight at 4°C in 4% paraformaldehyde (PFA), dehydrated in an increasing alcohol series, and CT-scanned using a Bruker SkyScan 1272 instrument (parameters: 70 kV, 142 µA, Al 0.5 mm filter, 13.2 µm resolution). For visualizing the dental lamina (DL), dissected jaws were fixed as above, dehydrated to 70% ethanol, and stained for 2–4 weeks with 0.3% phosphotungstic acid (PTA) in 70% ethanol, as described before (Metscher, 2009 (link)). Stained jaws were then CT-scanned in 70% ethanol (90 kV, 111 µA, Al 0.5 + Cu 0.038 mm filters, 2 µm resolution). All CT-scans were reconstructed using Bruker NRecon 1.7.0.4 software, and 3D volume rendering and segmentation of DL and/or teeth were done manually using Advanced 3D Visualization and Volume Modeling 5.5.0 (RRID:SCR_007353).

CT-scans of adult skulls (cranium and mandible) covering all major lineages of squamates (Additional Files 1, 2) were primarily obtained from the publicly available Digital Morphology Database (DigiMorph) or from our previous work (Da Silva et al., 2018 (link)). New high-resolution CT-scans of embryonic and adults skulls were produced at the University of Helsinki imaging facility using Skyscan 1272 microCT or Phoenix Nanotom 180, depending on the specimen size. To visualize skull bone development, fixed P. vitticeps embryos were CT-scanned at different days post-oviposition (15, 18, 24, 28, 32, 36, 40, 48, and 60 dpo) using the following parameters: filter: Al 0.25 mm; voltage: 60 kV; current: 166 μA; resolution: 6 μm. Scans were reconstructed using Bruker NRecon 1.7.0.4 software, and 3D isosurface rendering as well as segmentation of cranial bones were done using a variety of density thresholds with the software Amira 5.5.0 (Visualization Sciences Group). All 3D data were scaled by voxel size based on scan log file in Amira 5.5.0. For overall visualization of soft tissue and comparative morphology at early embryonic stages, embryos were first stained with 0.6% phosphotungstic acid (PTA) in ethanol for 14 days, as described before (Metscher, 2009 (link)), before scanning using the following parameters: filter: Al 1 mm; voltage: 80 kV; current: 125 μA; resolution: 6 μm.

High-resolution 3D CT-scans of P. vitticeps and B. fuliginosus heads were performed at the University of Helsinki imaging facility using Skyscan 1272 (Brucker, Belgium). Prior to micro-CT scanning, to allow reptile brain tissue visualization, samples were treated with 1% iodine solution as previously described (Metscher, 2009 (link); Macrì et al., 2019 (link)). The following scan parameters were used: filter: Al 0.25 mm; source voltage: 60 kV; source current: 166 μA; voxel size: 12 μm; rotation steps: 0,2°; frame averaging: 8. Scans were reconstructed using NRecon 1.7.0.4 software (Bruker) and 3D volume rendering as well as segmentation were performed using the software Amira 5.5.0 (Thermo Fisher Scientific, United States). Both the whole-brain and isolated cerebellum were segmented, thus allowing assessment of volumetric measurements of these structures. The accuracy of all generated 3D models was carefully controlled along the three anatomical planes, as described (Macrì et al., 2019 (link)).

The microstructure of the developed cookies was analyzed at different fill density values of 50%, 70%, 90%, and 100%. The printed cookies after baking were kept at room temperature for 2 h. The 3D images of the cookies were developed using a bench-top Micro-CT imager (SkyScan 1076, Bruker-MicroCT, Kontich, Belgium) at 18-µm voxel image resolution with 70 kV, 100 µA, and a 1-mm aluminum filter to remove low energy photons. Raw image projections of the cookies were reconstructed using a modified Feldkamp back-projection algorithm available with the bundled vendor software (Nrecon 1.6.1.5, Bruker-MicroCT, Kontich, Belgium). Representative region of interest (ROI) were sampled from the cross-sectional reconstructed images for all cookies and analyzed for structural parameters using vendor bundled CT-Analyze software (ver 1.11.6.0, Bruker-MicroCT, Kontich, Belgium). The same minimum and maximum threshold values were used to binarize each cookie in order to facilitate image analysis [16 (link),17 (link)]. The structural parameters determined using MicroCT analysis were open porosity, closed porosity, and total porosity. The parameters were calculated from three-dimensional analysis of each volume of interest (VOI) stack of binarized slices through the entire thickness of each cookie.

Front paws and tibias from mice were imaged using a Skyscan 1172 micro-CT scanner (Bruker) using X-ray beam settings of 60 kV/167 μA with a 0.5-mm aluminium filter. Projections were taken every 0.45° at 580-ms exposure. Image volumes were reconstructed using the Feldkamp algorithm (NRecon 1.6.1.5, Bruker) having applied beam hardening correction. Trabecular bone parameters (bone volume to tissue volume (BV/TV), trabecular thickness (Tb.Th) and trabecular number (Tb.N)) of the tibia were analysed using CTAn software. One millimetre of bone (150 sections) in the metaphyseal region beneath the growth plate was analysed, and regions of interest (ROI) were selected by drawing around the trabecular network for each cross-sectional slice. Front paws were reconstructed, and MeshLab 1.3.2 was used to generate meshes which could then be scored for bone erosions as described previously [18 (link)].

For micro-CT scanning one specimen each fixed in Bouin-solution of Polydesmusangustus (Latzel, 1884), Oxidusgracilis (Koch, C. L., 1847), Coromusvittatus (Cook, 1896) and Tymbodesmus sp. were transferred to 96% ethanol via an ascending ethanol series and stained with 3% Iodine solution for 24 hours. The specimens were washed in 100% ethanol and critical point dried using a Leica EM CPD 300. Micro-CT scanning was performed at the ZFMK using a SKYSCAN 1272 (Bruker micro-CT) with random movement = 15 and flat-filed correction and geometric correction switched on. For varying scanning parameters see Table 1. Post-alignment, ring-artefact reduction, beam-hardening correction and reconstruction were performed in NRecon 1.7.1.6 (Bruker microCT). The image stacks were modified using Fiji ImageJ 1.50e (Schindelin et al. 2012 (link)). Volume rendering was performed in Drishti Version 2.6.3 (Limaye 2012 (link)). Segmentation was done in ITK-SNAP 3.6.0 (Yushkevich et al. 2006 (link)). Images were edited in GIMP version 2.10.6 (https://www.gimp.org) and Inkscape 0.92 (www.inkscape.org).