Improving prokaryotic transposable elements identification using a combination of de novo and profile HMM methods

Table 2 Performance of a BLAST based search compared to a profile HMM search for diverse simulate and real metagenomes

Dataset	BLAST ISs false positives	HMM search ISs false positives
S. solfataricus, 1 kb	149 0	133 0
S. solfataricus, 250 bp	149 0	129 0
S. solfataricus, 100 bp	149 0	29 0
SAR11 Metagenome, ~250 bp	20 NA	13 NA
PBS Metagenome, ~1 kb	189 281	264 7
JCVI Metagenome, ~900 bp	44 114	87 0

Sulfolobus datasets correspond to in silico fragmented genomes and the results indicated the number of individual transposases identified. For the real marine metagenomes, the results indicate the number of apparent true transposases identified (false or ambiguous positives have been removed after careful alignments visualization) in addition to the numbers of false positives. NA: not applicable (see the text for further details).

ISSN: 1471-2164