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.