Although likely to exclude a number of bona fide targets, we used a maximum expectation value threshold of two in order to minimise false-positives.
The other phyla containing thermophiles or hyperthermophiles are shown in table 1. BlastP version 2.2.19 (Altschul et al. 1997) using a maximum expectation value threshold of 10−, and the inclusion of up to 100,000 hits per query, was used to compare the encoded proteins of all three Aquificae genomes against the full set of genomes.
Default settings were used as analytic parameters, i.e. 3 as the threshold of maximum expectation, 2 as the number of the multiplicity of target sites, and 9 to 11 nt as the range of the central mismatched region.
Maximum TPR values were obtained at the following threshold: psRNATarget: 5 (Maximum expectation), Tapirfasta: 10 (Score), Tapirhybrid: 10 (Score), psRobot: 5 (Penalty score threshold), Target_Prediction: 4 (Penalty score), Targetfinder: 6.5 (Prediction Score), Target-align: 110 (Maximum Score), p-TAREF: 4 (SVR score), RNAhybrid: 1 (p-value).
A maximum expectation value of 0.1 was employed as a threshold for peptide identifications.
A minimal ProteinProspector protein score of 15, a peptide score of 15, a maximum expectation value of 0.1, and a minimal discriminant score threshold of 0.0 were used for initial identification criteria.
Blast parameters used (WUBlast2.0): B 120 maximum number of database sequences for which any alignments will be reported E 1E-15 thexpectationon threshold for reporting database hits W 11 seed word length for the ungapped BLAST algorithm.
The different parameters are estimated by using the Maximum Expectation of joint Likelihood criterion.
We applied a strict expectation threshold of e = 0.01.
The maximum expectation value was set to 10-3, and a maximum of 25 alignments were reported per hit.
