心肌细胞钙离子荧光强度数据的数值处理方法比较

Comparison of Numerical Methods to Process Data of Myocytes Ca^(2+) Fluointersity

为比较多种用于减少非周期生理信号序列中的随机误差的线性叠加法 ,分析心房肌细胞的钙离子荧光强度满足的分布函数 ,提取特征参数 (斜率和时间常数 ) ,对 6个豚鼠心房肌细胞 ,建立 3种不同浓度的CGRP组和空白对照组 ,用共聚焦显微镜记录下 2 4组心房肌细胞的钙离子荧光强度数据 ;对这些数据进行校准、叠加、非线性拟合和提取特征参数。叠加方法包括等权平均叠加、主分量叠加、奇异值分解和最大似然叠加。选择Gamma分布函数作为拟合函数 ,使用最小二乘法进行拟合 ,并从Gamma函数计算特征参数。结果 :①给出 4种叠加方法的加权系数 ,4种线性叠加法的加权系数比较接近 ,特别是主分量法、奇异值分解法和最大似然法的加权系数几乎一样。对于主分量法 ,6个细胞的加权系数分别是 0 .1 6 2、0 .1 71、0 .1 6 7、0 .1 6 2、0 .1 70和 0 .1 6 8。②叠加后的荧光强度特征更加明显。③给出特征参数 (斜率和时间常数 )。用主分量叠加法得到的 3种不同浓度水平的CGRP组和空白对照组的斜率分别是1 2 77、1 32 6、2 0 39和 81 8,时间常数分别是 0 .33、0 .2 9、0 .1 9和 0 .2 7。结果提示 :对于只具有一个主要成分的实验数据 ,推荐使用主分量叠加法。这种方法比较简单 ,用它可以得到很好的叠加结果 。

Compare various linear superposition methods to reduce the random errors in the sequences of aperiodic physiological signals and analyze the distribution function adaptable to the myocytes Ca + fluointersity and draw the characteristic parameters-slopes and time constants. Methods: make the group of CRPR and the group of blank control for three different kinds of potency with six cells of cavy and note dawn the data of 24 groups of myocytes Ca + fluointersity with focusing microscope. Calibrate, superpose and make nonlinear fitting to these data and dawn the characteristic parameters from them.The methods of superposition include average superposition with equal weight, superposition of principal component, decomposition of the singular values and maximum likelihood superposition. Take the Gamma distribution function as the fitting function, fit with the way of least square and calculate the characteristic parameters from the Gamma function. Results: ① Give out the weighting coefficients from four superposition methods. The coefficients are approximate, especially the one from the method of principal component, decomposition singular values and maximum likelihood are nearly equal. In the method of principal component, the weighting coefficients of six cells are 0.162, 0.171, 0.167, 0.162, 0.170 and 0.168. ② The characteristics of the fluointersity after superposition is more clear. ③ Give out the characteristic parameters of slopes and time constants. The slopes of CGRP group and blank control group,which have three different kinds of potency and come from the superposition of principal component are 1 277 , 1 326, 2 039 and 818, while the time constants are 0.33, 0.29, 0.19 and 0.27. In the case of sample data with only one principal component, the superposition of principal component is recommended. This is a simple way. By this way, we can get a good result of superposition, fit the distribution very well and draw the characteristic parameters.