A novel pathway-based distance score enhances assessment of disease heterogeneity in gene expression

BMC Bioinformatics

Table 2 Phenotypic and physiologic characteristics of the identified clusters

	Euclid_all	Euclid_KEGG	KEGG_dist	Pathifier_KEGG
Age at Visit (years)	0.65	0.37	0.32	0.28
Gender	0.02*	0.14	0.58	0.28
History of Atopy - N (%)	0.89	0.2	0.02	0.62
Age of Symptom Onset	0.55	0.25	0.17	0.62
Disease Duration (years)	0.98	0.9	0.67	0.38
History of Hospitalization - N (%)	0.21	0.77	0.04	1.00
History of Intubations - N (%)	0.14	0.12	0.05	0.04
OCS tapers in past year- N (%)	0.65	1.00	0.67	0.83
ACT Score	0.25	0.41	0.22	0.56
FEV1- % of predicted value
Pre β ₂ agonist use	0.04	0.02	0.02	0.04
Post β₂ agonist use	0.06	0.05*	0.06	0.06
FVC- % of predicted value
Pre β ₂ agonist use	0.04	0.02	0.04	0.03
Post β₂ agonist use	0.12	0.06	0.16	0.13
FEV1/FVC- % of predicted value
Pre β₂ agonist use	0.23	0.46	0.13	0.41
Post β₂ agonist use	0.14	0.2	0.06	0.09
BDR (%)	0.27	0.05	0.05	0.09
FENO (ppb)	0.05*	0.54	0.27	0.40

The significance of the association between the phenotypic and physiologic features and the clustering results by K-means coupled with the four distances. P values were calculated using Kruskal-Wallis and Chi-square test for continuous and categorical variables, respectively. The false discovery rate for KEGG_dist clustering results associated clinical features estimated by the permutation-based method is 11% when nominal p value < 0.05. *P values that are significant (P < 0.05) only by one of the two Euclidean distances. Bold p values are significant by the pathway-based distance score

ISSN: 1471-2105