Inflammation and Immune System Activation in Aging: A Mathematical Approach

Nikas, Jason B.

Having employed three different and independent methods of statistical significance, namely, ROC curve analysis, fold change, and P-value, I was able to identify 36 genes that were the most significant in terms of differential expression. Fig. 1b depicts the results of K-Means clustering analysis based on the expression of the top 36 most significant genes. All K-Means clustering analysis results (with respect to both the housekeeping genes and the 36 most significant genes) are shown in Supplementary Table 2 . As can be seen in both Fig. 1b and Supplementary Table 2 , there is a clear separation of the two groups. Fig. 2 depicts the heat map that resulted by plotting the expression of those 36 genes for all 40 subjects (15 young and 25 old). As can be seen by the relative intensities, all of the 36 most significant genes are over-expressed (red color) in the case of the old subjects as compared with the case of the young subjects (blue color). The direction of the differential expression of those 36 genes also appears in Table 1 . Moreover, Fig. 3 provides a 3D representation of the differential expression of those 36 genes between the two groups in a surface-contour plot. classify the 40 subjects with a high accuracy. Such a model would be valuable in future studies of global gene expression analysis of post-mortem hippocampal tissue investigating biological and chronological aging. To that end, I randomly selected approximately 70% of the subjects [11/15 young subjects and 18/25 old subjects] for the development of the function (henceforward referred to as super variable), and I used the remaining subjects (4 young and 7 old) solely for the purpose of validating the super variable. Employing a general methodology that I have previously introduced 6,7 , I was able to generate a super variable (function) that, based on the input of 7 genes from the 36 most significant genes, was able to identify/classify accurately all but one of the old subjects {subject # 33 [64 yrs (F)]} [sensitivity 5 (24/25) 5 0.96] and all of the young subjects [specificity 5 (15/15) 5 1.00]. Those overall results of the performance of the F 1 super variable were obtained by combining the results from the development and the validation phases. According to the rank that appears in Table 1 , the seven genes that provide the input to the F 1 super variable are: C4A (C4B), ADORA3, MS4A7, BCL6, CD44, C3AR1, and HLA-DRB1. All of those seven genes are, in terms of biological function, either genes of inflammation or genes of immune system activation. Supplementary Fig. 1 shows the F 1 super variable function in relation to its 7 input gene variables. Fig. 4 Table 3 show the overall results of the F 1 super variable, i.e. the F 1 scores of all 40 subjects used in this study, as well as their respective classification. Fig. 4 and Supplementary Table 3 were created by combining the results from the development phase (the F 1 scores of all 29 subjects that were randomly selected and used exclusively for the development of the model) with the results from the validation phase (the F 1 scores of all 11 subjects that were randomly selected and used exclusively for testing purposes). The results of the F 1 super variable in the development phase are shown in Supplementary Fig. 2 and Supplementary Table 4 , whereas the results in the validation phase are shown in Supplementary Fig. 3 and Supplementary Table 5 .