Table 3.

Summary of total GO terms emerging during REMc. Unclustered data were enriched for 71 GO terms. The first round of clustering yielded four clusters and a total of 166 GO terms, indicating an increase in the detection of functionally related genes by clustering. However, 42 of the GO terms across the four clusters were redundant (reducing the total number of unique GO terms to 124, given in parentheses), and 15 terms (“missed previous”) associated with the unclustered data were not associated with one of the four clusters. The cluster ID name indicates the parent-child relationships for each clustering. Clustering was performed recursively until no new clusters were obtained. See supplemental material for list of GO terms for each cluster.

	Unclustered	Cluster ID	Rd 1 GT	Cluster ID	Rd 2 GT	Cluster ID	Rd 3 GT
	71	1.0	69	2.0-0	31
				2.0-1	61
		1.1	20	2.1-0	1
				2.1-1	15
				2.1-2	12
				2.1-3	8
				2.1-4	0
		1.2	56	2.2-0	41	3.2-0-0	21
						3.2-0-1	7
						3.2-0-2	12
				2.2-1	10
				2.2-2	17	3.2-2-0	5
						3.2-2-1	19
		1.3	21	2.3-0	0
				2.3-1	16
				2.3-2	0
				2.3-3	2
Unique GT (total)	71(71)		124(166)		144(214)		146(220)
New in round			68		39		11
Missed previous			15		19		9