Skip to main content

Table 1 Distribution of genes among a range of p-values identified by various analyses

From: Toxicogenomic analysis incorporating operon-transcriptional coupling and toxicant concentration-expression response: analysis of MX-treated Salmonella

 

Percent (%) of genes distributed across the range of p-values

Range of p-values

All genes in the genome*

All genes in selected operons†

Non-significant genes in selected operons‡

Genes identified by monotonic analysis§

0.0 – 0.2

15

72

37

94

0.2 – 0.4

17

11

24

6

0.4 – 0.6

20

6

14

0

0.6 – 0.8

23

5

12

0

0.8 – 1.0

25

6

13

0

  1. *p-values were calculated by CyberT comparisons of gene expression between the control cells and those treated with the highest concentration of MX (4.6 μM) for all 4253 genes analyzed. Thus, 4253 was the denominator used to calculate the percentages in this column. CyberT analysis identified 169 genes having p-values < 0.05.
  2. †Selected operons were those that had at least one gene with a consistent CyberT-calculated p-value < 0.05. Of the 169 genes identified by CyberT analysis, 86 resided in a total of 54 multi-gene operons, and the remaining 83 genes were in single-gene operons. Adding the additional genes on the 54 multi-gene operons (143) to those identified solely by CyberT (169) brought the total genes considered for analysis by the operon method to 312, which was the denominator used to calculate the percentages in this column. A Monte Carlo simulation, repeated 1500 times using a sample size of 312, randomly selected from 4253 p-values, and sorted into the p-value bins shown in column 1, produced a p-value distribution the same to the nearest percentage point as that shown in the second column, i.e., a distribution essentially identical to that found for all genes in the genome.
  3. ‡There were 161 genes residing in the selected operons that did not have p-values < 0.05. Eleven percent of these genes had missing data and, thus, did not appear in the table. Thus, 143 was the denominator used to calculate the percentages in this column. A Monte Carlo simulation repeated 1500 times using a sample size of 143 produced a p-value distribution the same to the nearest percentage point as that shown in the second column, i.e., a distribution essentially identical to that found for all genes in the genome.
  4. §The monotonic analysis alone identified 309 genes as being differentially expressed. Of these, 153 were uniquely identified by the monotonic analysis; the other 156 had been identified by either the CyberT and/or operon analyses. Thus, the monotonic analysis added an additional 153 genes for analysis, and this number was the denominator used to calculate the percentages shown in this column.
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us