Skip to main content
. 2020 Dec 14;27(6):870–886. doi: 10.1093/ibd/izaa281

TABLE 1.

Results of the Unsupervised Clustering of Samples Based on the Explanatory Latent Factors Associated with the Traits.

Trait LFa Dominantomic
layer
Correlation Between Cluster Membership
and Clinical Rraits
R2 P Optimal Clustering Enrichment of Traits in Clustersb
Disease location LF12 CD14 Yes 0.42 0.0164 k = 2 No
Disease behavior LF11 CD14 Yes −0.29 0.0965 k = 5 No
Disease behavior LF11 CD4 Yes −0.29 0.0965 k = 5 No
Perianal disease LF9 CD14 Yes 0.45 0.00834 k = 3 Yes
Number of anti-TNF agents exposed to LF5 CD14 Yes 0.437 0.0109 k = 2 No
Exposure to Vedolizumab LF3 CD14 Yes 0.563 0.0006 k = 2 Yes
Exposure to Vedolizumab LF5 CD14 Yes 0.46 0.007 k = 2 Yes
Smoking LF9 CD14 Yes −0.3255 0.6453 k = 2 No

aEnrichment was deemed as significant for those overrepresentation events with an adjusted P value of ≤ 0.1.

bLFs associated with gender were discarded to exclude sex specific signals.