Table 2: Median relative abundances of selected taxa according to smoking status in four data sets

From: Cigarette smoking and the oral microbiome in a large study of American adults

TaxaCPS-II-a (n=203)CPS-II-b (n=340)PLCO-a (n=261)PLCO-b (n=400)Meta P-valuea (q-valueb)Meta P-valuec (q-valued)
      Current vs NeverFormer vs Never
Proteobacteria (phylum)
 Never13.1813.4210.437.4   
 Former10.9712.557.597.21   
 Current7.152.633.624.28   
P-valuee0.020.0039.29E−060.0072.29E−09 (1.15E−08)1.05E−07 (5.24E−07)0.27 (0.60)
        
Betaproteobacteria (class); Proteobacteria (phylum)
 Never3.152.352.81.86   
 Former2.152.222.052.28   
 Current0.130.240.280.65   
P-valuee2.18E−055.07E−052.27E−080.00073.88E−16 (5.05E−15)1.02E−07 (1.02E−06)0.17 (0.92)
        
Gammaproteobacteria (class); Proteobacteria (phylum)
 Never8.938.525.234.01   
 Former7.028.53.893.96   
 Current6.341.872.472.38   
P-valuee0.170.030.0010.025.57E−06 (3.62E−05)5.73E−05 (0.0003)0.52 (0.92)
        
Flavobacteriia (class); Bacteroidetes (phylum)
 Never0.460.380.240.29   
 Former0.490.390.210.31   
 Current0.100.150.080.17   
P-valuee0.010.040.0010.336.11E−05 (0.0003)0.003 (0.009)0.31 (0.92)
        
Actinobacteria (phylum)
 Never10.5910.1015.1811.62   
 Former11.169.3013.8111.53   
 Current13.6513.8817.4611.55   
P-valuee0.350.620.360.710.47 (0.59)0.02 (0.04)0.87 (0.87)
        
Coriobacteriia (class); Actinobacteria (phylum)
 Never0.360.280.280.33   
 Former0.510.330.220.29   
 Current0.400.460.510.42   
P-valuee0.410.260.020.100.004 (0.01)0.01 (0.02)0.67 (0.92)
  1. aMeta-analysis P-values from Kruskal–Wallis tests within each of the four data sets, calculated using Z-score methods.
  2. bFalse discovery rate adjusted q-values were calculated based on the meta-analysis P-values from Kruskal–Wallis tests.
  3. cMeta-analysis P-values from polytomous logistic regression models within each of the four data sets, calculated using Z-score methods. Age (continuous value) and sex (male, female) were controlled for in the polytomous regression models.
  4. dFalse discovery rate adjusted q-values were calculated based on the meta-analysis P-values from polytomous logistic regression models.
  5. eP-values are based on nonparametric Kruskal–Wallis test of current, former and never smokers.