Dataset from our paper entitled "Predicting Spatial Metabolic Profiles: How Cell Morphology Impacts Energy Metabolism"

Dataset from our paper entitled "Predicting Spatial Metabolic Profiles: How Cell Morphology Impacts Energy Metabolism" This repo contains a ‘minimal data set’ necessary to interpret, replicate and build on the findings reported our paper entitled "Predicting Spatial Metabolic Profiles: How Cell Morphology Impacts Energy Metabolism". This dataset was produced by a computational model developed by Dr. Sofia Farina and presented in detail here: Farina, S., Claus, S., Hale, J.S. et al. A cut finite element method for spatially resolved energy metabolism models in complex neuro-cell morphologies with minimal remeshing. Adv. Model. and Simul. in Eng. Sci. 8, 5 (2021). https://doi.org/10.1186/s40323-021-00191-8 Please find below some comments explaining the organization of the data. For the concentration files Each row corresponds to a realization. The concentrations are organized as follows: Each 6 sequential columns correspond to one grid point (the grid points consisting of x and y coordinates are included in the file grid.csv). The metabolite concentrations are ordered as follows: [GLC], [ATP], [ADP], [GLY], [PYR], [LAC] This means that the first 6 columns correspond to the first grid point. The first column corresponds to the [GLC] concentration, the second one the [ATP] concentration, the third one to the [ADP] concentration etc. For the reaction site files There are 40 reaction centers for each realization. Each center has a set of (x,y) coordinates for a total of 80 cordinates per realization. The data files are organized as follows: The realization column indicates the realization number. For each realization number there is a total of 80 coordinates, organized in 10 rows and 8 columns (x_hxk,y_hxk,x_pyrk,y_pyrk,x_ldh,y_ldh,x_mito,y_mito).本数据集来自题为《基于机器学习的空间代谢谱预测揭示形态对星形胶质细胞能量代谢的影响》的研究论文。 本代码仓库包含该论文中部分具有代表性的数据集。本数据集由索菲亚·法里纳（Sofia Farina）博士开发的计算模型生成，相关细节已发表于以下文献：Farina, S., Claus, S., Hale, J.S. 等. 面向复杂神经细胞形态下空间分辨能量代谢模型的最小重网格化切割有限元方法[J]. 工程科学建模与仿真进展, 8, 5 (2021). https://doi.org/10.1186/s40323-021-00191-8 下文将对本数据集的组织形式进行说明。 针对浓度数据文件： 每一行对应一次仿真实现。浓度数据的组织形式如下： 每连续6列对应一个网格节点（包含x、y坐标的网格节点数据已单独存入grid.csv文件中）。 代谢物浓度的排序依次为：葡萄糖（GLC）、三磷酸腺苷（ATP）、二磷酸腺苷（ADP）、甘油（GLY）、丙酮酸（PYR）、乳酸（LAC）。 这意味着前6列对应第一个网格节点。其中第一列为葡萄糖（GLC）浓度，第二列为三磷酸腺苷（ATP）浓度，第三列为二磷酸腺苷（ADP）浓度，以此类推。 针对反应位点数据文件： 每个仿真实现包含40个反应中心，每个反应中心带有一组(x,y)坐标，因此每个仿真实现总计包含80个坐标值。数据文件的组织形式如下： 第一列为仿真实现编号，用于标识对应的算例。 每个仿真实现对应总计80个坐标值，以10行8列的格式存储，各列依次为：己糖激酶（hexokinase, hxk）x坐标、己糖激酶（hexokinase, hxk）y坐标、丙酮酸激酶（pyruvate kinase, pyrk）x坐标、丙酮酸激酶（pyruvate kinase, pyrk）y坐标、乳酸脱氢酶（lactate dehydrogenase, ldh）x坐标、乳酸脱氢酶（lactate dehydrogenase, ldh）y坐标、线粒体（mitochondrion, mito）x坐标、线粒体（mitochondrion, mito）y坐标。