Table 5.

Wall-clock running time (hr) for UPP (Fast) and UPP (Default) for the RNASim 10K and CRW 16S.B.ALL datasets

	RNASim 10K		CRW 16S.B.ALL
Stage	UPP (Fast)	UPP (Default)	UPP (Fast)	UPP (Default)
Building backbone	0.12	0.42	0.13	0.52
Building HMMs	<0.01	0.02	<0.01	0.02
Searching for best HMM	0.83	6.53	1.81	15.45
Aligning sequences	0.02	0.03	0.05	0.15
Merge alignments	0.01	0.01	0.01	0.02
Total time	0.99	7.01	2.01	16.16

Wall-clock running time (hr) for each stage of UPP (Fast) and UPP (Default) for the RNASim 10K (10,000 sequences) and CRW 16S.B.ALL (27,643 sequences) datasets, two of the largest nucleotide datasets. The UPP alignments were computed on TACC’s Lonestar Cluster machine. The vast majority of the running time was spent searching for the best HMM for the query sequences.