The influence of age on REOs within normal lung tissues
From three datasets (GSE31210, GSE19804 and GSE20257, as shown in Table 5), we selected 65 samples of non-smoking Asian females with age ranging from 37 to 80 years old for the analysis. Based on the REO pattern of each gene pair, the samples were divided into two groups, and then the Mann-Whitney U-test was used to test whether there is significant difference in age between the two groups of samples. We could not find any gene pair whose REO was significantly correlated with age with FDR < 0.05 or even with FDR < 0.2 (Methods). Similarly, using 34 samples for Caucasian males with age ranging from 27 to 80 years old, collected from the dataset GSE4115, no significant gene pair was found with either FDR < 0.05 or FDR < 0.2.
The above results indicated that the influence of age on REO of gene pair could be negligible. Accordingly, the age factor was not considered in the subsequent analyses.
The influence of cigarette smoking on REOs within normal lung tissues
We compared the gene expression profiles of normal lung tissue samples for 49 smokers and 44 non-smokers from the GSE20257 dataset. The detailed information on the sample composition was shown in Table 1. There is no significant difference in sex or race distribution between the smoker group and the non-smoker group (Fisher’s exact test, P > 0.1).
With FDR < 0.05, we identified the gene pairs with significantly stable REOs in the smoker group and non-smoker group, respectively. We found 187,875,560 gene pairs that have significantly stable REOs (binomial test, FDR < 0.05) in both groups, among which 0.227% showed reversal REO patterns. With RankCompV2, we identified 344 DEGs, including 210 up- and 134 down-regulated genes in the smoker group compared with the non-smoker group (FDR < 0.05). The 210 up-regulated genes and 134 down-regulated genes were enriched, respectively, in 7 pathways and 1 pathway (hypergeometric test, FDR < 0.05), as shown in Figure 1. For the pathway “metabolism of xenobiotics by cytochrome P450”, cytochrome P450 are known to be responsible for the metabolism of compounds present in cigarette smoke, including nicotine, benzene, polycyclic aromatic hydrocarbons (PAHs) and tobacco-specific nitrosamines (TSNAs)[30]. As for the “glutathione metabolism” pathway, it has been found that cigarette smoking could induce the deregulation of glutathione metabolism in bronchial epithelial cells[31]. It has also been reported that “metabolic pathways”[32], “steroid hormone biosynthesis”[33], “pentose phosphate pathway”[34], “arachidonic acid metabolism”[35] and “mineral absorption”[36] are affected by cigarette smoking.
The above results indicated that cigarette smoking can alter the REOs in normal lung tissues and disturb some important biological pathways.
The influence of sex on REOs within normal lung tissues
We compared the gene expression profiles of normal lung tissue samples for 64 males and 29 females from the dataset GSE20257. The detailed information of the sample composition was shown in Table 2. There is no significant difference in smoking rate or race distribution between the male group and the female group (Fisher’s exact test, P > 0.2).
We identified the gene pairs with significantly stable REOs in the male and female groups, respectively, and found 187,481,246 gene pairs with significantly stable REOs (binomial test, FDR < 0.05) in both groups, among which 0.074% showed the reversal REO patterns. With RankCompV2, we identified 35 DEGs in the male group compared with the female group (FDR < 0.05). In another dataset GSE71181, including 201 male samples and 80 female samples which are all from smokers, 25 of the above 35 DEGs were also found (T-test, FDR < 0.05) and 96% (24 genes) have the same dysregulation directions in the male group compared with the female group. Among the 24 DEGs, 6 out of the 10 up-regulated genes in the male group are located on Y chromosome, 12 out of the 14 up-regulated genes in the female group are located on X chromosome, and the cytoband of these genes is shown in Table 3. In particular, DDX43, CRISP2 and PRDM7, which are up-regulated in the male group are located on autosome and involved in spermatogenesis and male fertility[37, 38]. For the other two genes, NLRP2 and C3orf79, located on autosome but up-regulated in the females, it is known that NLRP2 is a critical regulator of oocyte[39].
The influence of the race factor on REOs within normal lung tissues
Due to the limitation of the sample sizes for other races, we only compared the gene expression profiles of normal lung tissues for the white and black races. From the GSE20257 dataset, we obtained 34 samples for white people and 59 samples for black people. The detailed information of the sample composition was shown in Table 4. There is no significant difference in cigarette smoking rate or sex distribution between the two groups (Fisher’s exact test, P > 0.1).
With FDR < 0.05, we found 187,973,147 gene pairs with significantly stable REOs in both groups, among which 0.0272% showed reversal REO patterns. With RankCompV2, we identified 22 DEGs, including 10 up- and 12 down-regulated genes in the white group compared with the black group (FDR < 0.05). Due to the small number of DEGs, we found no pathway significantly enriched with the up- or down-regulated DEGs with FDR < 0.05. With P < 0.05, the 10 up-regulated and 12 down-regulated genes were enriched in, respectively, 4 and 4 pathways, as shown in Figure 2. The result indicates that there are some differences in metabolism and immunity of the normal lung tissues between the white and black races[40, 41].