Skip to main content

Advertisement

Table 7 Diabetes dataset covariates identified as correlated with principal components. Results of a search for covariates with value distributions having significant differences in either UP n vs. DOWN n , {UP n +FLAT n } vs. DOWN n , or UP n vs. {FLAT n +DOWN n } are shown below. An "X" indicates the covariate's values varied significantly in the corresponding principal component's sample partitioning. 11 (out of 53) covariates were identified at a 1% significance level (minSetSize = 5) in 10 of the first 20 principal components. To investigate further, corresponding plots of covariate value distributions within sample partitions can be generated, such as the covariate plots for PC14 (see Figure 5).

From: Mining gene expression data by interpreting principal components

  Principal Component Number
  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Cap_(mm2)   X                   
Cap__Type2b_(mean_n)            X          
Centroid_(Using_34_OXPHOS-CR_Genes)         X             
Insulin_0               X       
Patient_#     X                 
Type1_Min_Area_(um2)          X            
Type2a_(n)        X              
Type2a_Min_Area_(um2)                 X     
Type2b_(%)               X       
Type2b_(n)    X                  
UQCRB_(209065_at)       X