Integrated Weighted Gene Co-expression Network Analysis with an Application to Chronic Fatigue Syndrome

BMC Systems Biology

Table 2 Understanding the factors that affect gene significance.

Data Set	No. Samples	a.) cor(severity, MEblue)¹	b.) cor(severity, SNP)²
Samples with severity scores	87	r = 0.27 (p = 0.011)	r = 0.28 (p = 0.010)
Males & HomFemales	76	r = 0.50 (p = 8 × 10-6)	r = 0.33 (p = 0.003)
Males with severity scores	23	r = 0.34 (p = 0.113)	r = 0.45 (p = 0.030)
Females with severity scores	64	r = 0.26 (p = 0.041)	r = 0.17 (p = 0.170)
Homogenized Females	53	r = 0.54 (p = 2 × 10-5)	r = 0.25 (p = 0.076)
Second data set (DS)³	33	r = 0.09 (p = 0.638)	r = 0.03 (p = 0.846)
Second DS Homogenized³	30	r = 0.38 (p = 0.040)	r = 0.15 (p = 0.415)
Data Set	No. Samples	c.) cor(severity, *FOXN1* )	d.) cor(SNP, *FOXN1* )
Samples with severity scores	87	r = 0.21 (p = 0.055)	r = 0.18 (p = 0.088)
Males & HomFemales	76	r = 0.38 (p = 6 × 10-4)	r = 0.23 (p = 0.045)
Males with severity scores	23	r = 0.27 (p = 0.216)	r = 0.20 (p = 0.365)
Females with severity scores	64	r = 0.21 (p = 0.101)	r = 0.19 (p = 0.137)
Homogenized Females	53	r = 0.44 (p = 0.001)	r = 0.24 (p = 0.082)
Second DS³	33	r = 0.28 (p = 0.116)	r = 0.18 (p = 0.318)
Second DS Homogenized³	30	r = 0.40 (p = 0.030)	r = 0.12 (p = 0.515)

The severity, MEblue, TPH2 SNP, and FOXN1 correlations (r) and p-values (p) for five different subsets of the first primary data set and the homogenized samples in the secondary data set. Severity is most significantly related toMEblue, the TPH2 SNP and the candidate FOXN1 gene in the combined male and homogenized female sub set. Homogenizing the second data set (which discarded three samples) also improved the MEblue and FOXN1 associations with severity.
¹MEblue refers to the blue module eigengene, or first principal component of the blue module.
²SNP refers to the TPH2 SNP rs10784941.
³Here empiric replaces severity in cor(severity, FOXN1).

ISSN: 1752-0509