Genome-wide Terminator Prediction

Genomic data for the 343 organisms was downloaded on 6 June 2006 from [25 ]. Annotations were taken from the .ptt file accompanying the genomic sequence. Accession and version numbers for the sequences used are available as Additional data file 3. TransTermHP was run using the default search parameters and the -p scoring parameter to produce predictions for the best terminator following each gene in these organisms (available on the web [18 ], --bag option to TransTermHP) as well as the best terminators in each robust tail-to-tail region (Additional data file 2, --t2t-perf option to TransTermHP). The former were used when comparing performance to [3 (link)], while the latter was used to assess the importance of Rho-independent termination within an organism. Experiments were run on a 3.2 GHz Intel Xeon processor running Linux. High-confidence terminators are those with scores ≥ 76.
The C++ source code for TransTermHP is available (Additional data file 1) under an open source license.