Data from: Integrative approach uncovers new patterns of eco-morphological convergence in slow arboreal xenarthrans

<b>A) Alfieri et al. 2021_statistical analysis.R</b>: code allowing to replicate all the steps of the statistical analysis. All the functions are applied, as example, on humeral data (starting from SOM1_humeri.xls). The same steps can be applied to femoral data. The file also include the univariate version of convevol functions. <b>B) Choloepus didactylus ZMB Mam-102636_humerus_TEMP.ply and Bradypus sp ZMB Mam-33806_femur_TEMP.ply</b>: humeral and femoral templates which were subjected to inflation and decimation. Details and operations in which these models are involved are explained in the text <b>C) humeri3DGM_rep_array</b> and <b>femora3DGM_rep_array</b> : R workspaces to be loaded to analyse the arrays 'sorted', including landmark coordinates needed for the repeatability test on humeral and femoral external shape data <b>D) MacroSphereRect.ijm</b>: Fiji macro allowing to extract a spherical Region of Interest (ROI) of trabecular bone from an articular joint, as detailed in the text and in the macro <b>E) SlicebySliceCSP.ijm</b>: modified version of the macro from Amson, 2019 (details in the text), allowing to automatically compute non-directional cross-sectional properties with a slice-by-slice approach on thresholded, purified and oriented CTstacks. <b>F) SOM1_humeri.xls</b> and <b>SOM2_femora.xls</b>: complete datasets and information for the all the analysed humeri and femora. For each of them they include 1. Catalogue number, collection and taxon 2. CTscanning resolution 3. Metric measurements taken on the proximal[ <i>(HeadPDLen(mm), HeadMLLen(mm), HeadAPLen(mm)</i>] and distal [<i>CapPDLen(mm), CapMLLen(mm), CapAPLen(mm)</i> for the humeri;<i>LatConPDLen(mm), LatConMLLen(mm), LatConAPLen(mm) MedConPDLen(mm), MedConMLLen(mm), MedConAPLen(mm)</i> for the femora] epiphyses. 4. If the specimen was discarded from 3D GM analysis, <i>DiscGM=1</i> 5. If the specimen requested the process of correction for biased slices during the computation of CSP, <i>SliRemCSG=1</i> 6. If the specimen was too filled with sediment or too damaged and the 50% level was not preserved, <i>DiscCSG=1</i> 7. If extreme diaphyseal slices were biased, needing manual restoration to allow the data interpolation (as explained in the text),<i>ExtSliRest=1</i> 8. the specimen was too filled with sediment or too damaged but the 50% level was preserved and analysed, <i>Only50=1</i> 9. If the 50% level (preserved as in point 8) needed manual restoration to be analysed, <i>Rest50=1</i> 10. <i>DiscProxTrab</i> and <i>DiscDistTrab</i> for the humerus, are =1 if the extracted ROIs of trabecular bone were discarded for the proximal and/or distal epiphyses, respectively. <i>DiscProxTrab</i>, <i>DiscLatConTrab</i> and <i>DiscMedConTrab</i>, for the femur, are =1 if the extracted ROIs of trabecular bone were discarded for the proximal epiphysis, lateral condyle and/or medial condyle, respectively. 11. <i>RestProxTrab</i> and <i>RestDistTrab</i> for the humerus, are =1 if the extracted ROIs of trabecular bone from the proximal and/or distal epiphyses, respectively, needed restoration in order to be analysed. <i>RestProxTrab</i>, <i>RestLatConTrab</i> and <i>RestMedConTrab</i>, for the femur, are =1 if the extracted ROIs of trabecular bone from the proximal epiphysis, lateral condyle and/or medial condyle, respectively, needed restoration in order to be analysed. 12. Diameters of the extracted trabecular bone ROIs (<i>ROIproxDiam (mm) </i>and <i>ROIdistDiam (mm)</i> for the humerus; <i>ROIproxDiam (mm), ROILatConDiam (mm), ROIMedConDiam (mm)</i>; for the femur) 13. Relative resolution for each trabecular bone (computed and interpreted as detailed in the text); <i>ROIproxRelRes</i> and <i>ROIdistRelRes</i> for the humerus; <i>ROIproxRelRes</i>, <i>ROILatConRelRes</i> and <i>ROIMedConRelRes</i> for the femur; 14. Body mass proxy (BMproxy), computed as detailed in the text and in Supplementary Information 1 15. <i>PC1-PC38</i> (for the humerus) and <i>PC1-PC39</i> (for the femur): PC scores extracted from the 3D analysis 16. <i>ResC50 (%)- CSSAver</i>: CSP raw data, extracted from both the 50% level (Parameter50) and averaging CSP along the diaphysis (ParameterAver), as detailed in the text 17. <i>DAprox- Av.Br.LenDist(mm)</i> for the humerus; <i>DAprox-Av.Br.MedCon(mm)</i> for the femur: trabecular parameters raw data from each of the analysed articular joint (as detailed in the text). <b>G) SOM3-SOM13.xls</b>: raw data from three repeated data extractions on a subset of 7 humeri and femora, to assess repeatability (detailed in the text) <b>H) SuppInfo1_AdditionalMethods.pdf</b> <b>I) SuppInfo2_AdditionalResults.pdf</b> <b>J) xenarthra_tree.nex</b>: time-calibrated phylogenetic tree used in this study<br><br>

<b>A) Alfieri et al. 2021_statistical analysis.R</b>: 可复现全部统计分析步骤的代码。所有函数均以肱骨数据集（源自SOM1_humeri.xls）为例进行演示，该分析流程同样适用于股骨数据集。本文件同时包含convevol函数的单变量版本。<br><br><b>B) Choloepus didactylus ZMB Mam-102636_humerus_TEMP.ply 与 Bradypus sp ZMB Mam-33806_femur_TEMP.ply</b>: 经过膨胀与降采样处理的肱骨与股骨模板，相关模型的应用细节与操作步骤详见正文。<br><br><b>C) humeri3DGM_rep_array</b> 与 <b>femora3DGM_rep_array</b>: 用于加载并分析"sorted"数组的R工作区，其中包含肱骨与股骨外形可重复性检验所需的标志点坐标。<br><br><b>D) MacroSphereRect.ijm</b>: Fiji宏脚本，可从关节面提取松质骨的球形感兴趣区域（Region of Interest，简称ROI），具体操作详见正文与该宏脚本的说明文档。<br><br><b>E) SlicebySliceCSP.ijm</b>: 基于Amson 2019年脚本修改的宏版本，可针对经过阈值分割、图像纯化与轴向定向的CT堆叠数据（CT stack），通过逐切片方法自动计算无方向横截面特性（Cross-Sectional Properties，简称CSP），相关细节详见正文。<br><br><b>F) SOM1_humeri.xls</b> 与 <b>SOM2_femora.xls</b>: 本研究全部分析用肱骨与股骨的完整数据集及相关附属信息，每份数据集包含以下17项内容：<br>1. 标本编号、馆藏机构与分类类群<br>2. CT扫描分辨率<br>3. 分别于近端骨骺与远端骨骺采集的形态测量指标：针对肱骨，近端骨骺的测量参数为<t>HeadPDLen(mm)、HeadMLLen(mm)、HeadAPLen(mm)</t>，远端骨骺为<t>CapPDLen(mm)、CapMLLen(mm)、CapAPLen(mm)</t>；针对股骨，相关骨骺的测量参数为<t>LatConPDLen(mm)、LatConMLLen(mm)、LatConAPLen(mm)、MedConPDLen(mm)、MedConMLLen(mm)、MedConAPLen(mm)</t><br>4. 若标本未纳入三维几何形态测量（3D GM）分析，则标记<t>DiscGM=1</t><br>5. 若标本在横截面特性（CSP）计算过程中需对偏移切片进行校正，则标记<t>SliRemCSG=1</t><br>6. 若标本填充过多沉积物或损坏严重，无法保留50%截面水平，则标记<t>DiscCSG=1</t><br>7. 若骨干中段存在偏移切片，需手动修复以完成数据插值（详见正文），则标记<t>ExtSliRest=1</t><br>8. 若标本填充过多沉积物或损坏严重，但仍保留50%截面水平并纳入分析，则标记<t>Only50=1</t><br>9. 若保留的50%截面水平（对应第8项）需手动修复后方可分析，则标记<t>Rest50=1</t><br>10. 针对肱骨，<t>DiscProxTrab</t>与<t>DiscDistTrab</t>分别标记近端和/或远端骨骺的松质骨ROI是否被弃用，若弃用则取值为1；针对股骨，<t>DiscProxTrab</t>、<t>DiscLatConTrab</t>与<t>DiscMedConTrab</t>分别标记近端骨骺、外侧髁和/或内侧髁的松质骨ROI是否被弃用，若弃用则取值为1<br>11. 针对肱骨，<t>RestProxTrab</t>与<t>RestDistTrab</t>分别标记近端和/或远端骨骺的松质骨ROI是否需修复后方可分析，若需修复则取值为1；针对股骨，<t>RestProxTrab</t>、<t>RestLatConTrab</t>与<t>RestMedConTrab</t>分别标记近端骨骺、外侧髁和/或内侧髁的松质骨ROI是否需修复后方可分析，若需修复则取值为1<br>12. 提取的松质骨ROI直径：肱骨为<t>ROIproxDiam (mm)</t>与<t>ROIdistDiam (mm)</t>；股骨为<t>ROIproxDiam (mm)</t>、<t>ROILatConDiam (mm)</t>与<t>ROIMedConDiam (mm)</t><br>13. 每份松质骨样本的相对分辨率（计算与解读方式详见正文）：肱骨为<t>ROIproxRelRes</t>与<t>ROIdistRelRes</t>；股骨为<t>ROIproxRelRes</t>、<t>ROILatConRelRes</t>与<t>ROIMedConRelRes</t><br>14. 体质量代理变量（BMproxy），计算方式详见正文与补充信息1<br>15. 主成分分析得分：肱骨为<t>PC1–PC38</t>，股骨为<t>PC1–PC39</t>，均提取自三维几何形态测量分析<br>16. <t>ResC50 (%)–CSSAver</t>: 横截面特性原始数据，同时包含50%截面水平参数（Parameter50）与沿骨干平均的横截面特性参数（ParameterAver），具体细节详见正文<br>17. 针对肱骨为<t>DAprox–Av.Br.LenDist(mm)</t>，针对股骨为<t>DAprox–Av.Br.MedCon(mm)</t>: 各分析关节面的松质骨参数原始数据，具体细节详见正文<br><br><b>G) SOM3-SOM13.xls</b>: 来自7份肱骨与股骨样本子集的三次重复数据提取的原始数据集，用于评估分析可重复性（详见正文）<br><br><b>H) SuppInfo1_AdditionalMethods.pdf</b><br><br><b>I) SuppInfo2_AdditionalResults.pdf</b><br><br><b>J) xenarthra_tree.nex</b>: 本研究使用的时间校准系统发育树。