Finite element simulation data for the computation of the Young's modulus in real-time deformability cytometry

<strong>Finite element simulation data for the computation of the Young's modulus in real-time deformability cytometry<br> </strong><br> <strong>Background</strong><br> In real-time deformability cytometry (RT-DC), the area and the deformation of a sphere in a microfluidic channel at high flow rates can be translated into Young's modulus [1] by employing a look-up-table (LUT). For the average RT-DC user, these computations are usually performed by dclab [2].<br> <br> The refurbished LUT introduced in dclab 0.24.0 was omputed from this dataset. The original LUT in dclab (version &lt;=0.23.0) was slightly smaller and it was not possible to derive volume-deformation isoelasticity lines (because those data were not available). Dominic Mokbel and Lucas Wittwer reran all the original simulations (in addition to a few new ones) to obtain this dataset.<br> <br> <br> <strong>Dataset<br> </strong>The dataset "20191113_ObjectShapes_RT-DC_2Daxis.hdf5" contains object shapes obtained with finite element method (FEM) simulations which were originally performed by Mokbel et al. [3]. The simulations were done for a linear elastic material. The HDF5 file contains the final object shapes and (via attributes) the corresponding values for Young's modulus, deformation, projected area, and volume. The hdf5-File contains for each Young's modulus a group which has for each sphere radius a subgroup (numbered with i, j = 0, 1, 2, ...).<br> <br> <br> <strong>Notes<br> </strong>- The sphere is modeled as a linear elastic material.<br> - In RT-DC, objects are flushed through a rectangular channel. The FEM simulations, however, are performed in cylindrical symmetry (axis-symmetric). The equivalent channel radius for a square channel was computed with a factor of 1.094 and employed in the simulations (see [1] supplemt S3).<br> - The fact that the simulations were performed with cylindrical symmetry is also important for objects with a large cross-sectional area. Here, the flow profile in the cylindrical channel does not anymore correctly approximate the corresponding profile in the square channel, leading to inaccurate object shapes. Furthermore, there have been numerical errors due to meshing if the area is above 290um^2. Therefore, simulations with cross-sectional areas above 290um^2 should not be considered reliable (they are excluded from the LUT in dclab).<br> - There are no simulations below an area of 200um^2 and a deformation below approximately 0.005. This is because those deformation values were not resolvable with FEM. In dclab, those missing values are filled in with values computed using the analytical solution [1]. For the sake of completeness, these missing values are included here in the file "LUT_analytical_linear-elastic_2Daxis.txt".<br> - Compared to the initial LUT in dclab&lt;=0.23.0, the values given here differ by a relative error below 0.001 (0.1%).<br> - For more information, please refer to the data processing scripts in the dclab repository [4].<br> <br> <br> <strong>References<br> </strong>[1] Alexander Mietke, Oliver Otto, Salvatore Girardo, Philipp Rosendahl, Anna Taubenberger, Stefan Golfier, Elke Ulbricht, Sebastian Aland, Jochen Guck, and Elisabeth Fischer-Friedrich. Extracting Cell Stiffness from Real-Time Deformability Cytometry: Theory and Experiment. Biophysical Journal, 109(10):2023–2036, nov 2015. doi:10.1016/j.bpj.2015.09.006.<br> [2] https://dclab.readthedocs.io<br> [3] M. Mokbel, D. Mokbel, A. Mietke, N. Träber, S. Girardo, O. Otto, J. Guck, and S. Aland. Numerical Simulation of Real-Time Deformability Cytometry To Extract Cell Mechanical Properties. ACS Biomaterials Science &amp; Engineering, (11):2962–2973, jan 2017. doi:10.1021/acsbiomaterials.6b00558.<br> [4] https://github.com/ZELLMECHANIK-DRESDEN/dclab/tree/master/scripts<br> <br> <br> <strong>SHA256 sums<br> </strong>e41f5ccf0a13da7627791425bbd1efd8454093064a0e225e9d6339961b594ffc 20191113_ObjectShapes_RT-DC_2Daxis.hdf5<br> 908ded1a5bee8a39104eaef56d74ff692ab9a5c52777164af5ed4bf183a0cbe0 LUT_analytical_linear-elastic_2Daxis.txt<br> <br>

实时变形细胞术（real-time deformability cytometry, RT-DC）杨氏模量计算用有限元模拟数据集<br> <strong>背景</strong><br> 在实时变形细胞术（RT-DC）中，可通过查找表（look-up-table, LUT）将高流速下微流道内球体的投影面积与形变量转换为杨氏模量[1]。对于普通RT-DC使用者而言，这类计算通常由dclab工具[2]完成。<br> <br> dclab 0.24.0版本引入的更新版查找表正是基于本数据集生成。dclab早期版本（≤0.23.0）中的原始查找表尺寸偏小，且无法推导体积-形变等弹性线，因当时缺乏对应数据。Dominic Mokbel与Lucas Wittwer重新运行了全部原有模拟（并新增了少量模拟），最终构建了本数据集。<br> <br> <strong>数据集</strong><br> 本数据集包含的"20191113_ObjectShapes_RT-DC_2Daxis.hdf5"文件存储了由Mokbel等人[3]最初开展的有限元法（finite element method, FEM）模拟得到的物体形貌。本次模拟针对线性弹性材料展开。该HDF5文件不仅存储了最终的物体形貌，还通过属性字段附带了杨氏模量、形变量、投影面积与体积的对应数值。文件中按杨氏模量分组，每组下再按球体半径设置子组（编号规则为i, j = 0, 1, 2, ...）。<br> <br> <strong>说明</strong><br> - 球体被建模为线性弹性材料。<br> - 在RT-DC实验中，样品颗粒被冲入矩形流道内，但本次FEM模拟采用了圆柱对称（轴对称）建模。通过系数1.094计算得到方形流道的等效流道半径，并将其应用于模拟中（详见文献[1]补充材料S3）。<br> - 采用圆柱对称建模这一设定对于大横截面积的颗粒同样需要注意：此时圆柱流道内的流速分布无法准确匹配方形流道的流速分布，会导致物体形貌计算出现偏差。此外，当横截面积超过290μm²时，网格划分过程会引入数值误差。因此，横截面积大于290μm²的模拟结果不具备可靠性（该部分数据已从dclab的查找表中剔除）。<br> - 本数据集未包含横截面积小于200μm²或形变量约小于0.005的模拟结果，因该类形变量无法通过FEM解析。在dclab中，此类缺失值通过解析解[1]计算得到。为保证数据集完整性，补充文件"LUT_analytical_linear-elastic_2Daxis.txt"中包含了所有缺失的数值。<br> - 与dclab≤0.23.0版本中的原始查找表相比，本数据集给出的数值相对误差低于0.001（即0.1%）。<br> - 如需更多信息，请参考dclab代码仓库[4]中的数据处理脚本。<br> <br> <strong>参考文献</strong><br> [1] Alexander Mietke, Oliver Otto, Salvatore Girardo, Philipp Rosendahl, Anna Taubenberger, Stefan Golfier, Elke Ulbricht, Sebastian Aland, Jochen Guck, and Elisabeth Fischer-Friedrich. 从实时变形细胞术中提取细胞刚度：理论与实验. 生物物理期刊, 109(10):2023–2036, nov 2015. doi:10.1016/j.bpj.2015.09.006.<br> [2] https://dclab.readthedocs.io<br> [3] M. Mokbel, D. Mokbel, A. Mietke, N. Träber, S. Girardo, O. Otto, J. Guck, and S. Aland. 实时变形细胞术的数值模拟以提取细胞力学特性. ACS生物材料科学与工程, (11):2962–2973, jan 2017. doi:10.1021/acsbiomaterials.6b00558.<br> [4] https://github.com/ZELLMECHANIK-DRESDEN/dclab/tree/master/scripts<br> <br> <strong>SHA256校验和</strong><br> e41f5ccf0a13da7627791425bbd1efd8454093064a0e225e9d6339961b594ffc 20191113_ObjectShapes_RT-DC_2Daxis.hdf5<br> 908ded1a5bee8a39104eaef56d74ff692ab9a5c52777164af5ed4bf183a0cbe0 LUT_analytical_linear-elastic_2Daxis.txt