Gene prediction using the Self-Organizing Map: automatic generation of multiple gene models

BMC Bioinformatics

Table 1 Accuracy of RescueNet in 15 bacterial genomes.

Organism	GC %	Number of Genes Annotated	Training Set Size	Sn. (%)	Sn. >225 bp (%)	Sn. Conserved (%)	Sp. (%)
*Buchnera*	26.2	564	292	88.65	91.24	89.97	96.18
*B. burgdorferi*	28.6	857	403	90.54	96.39	95.66	98.02
*C. jejuni*	30.6	1654	673	90.14	95.08	92.14	99.23
*M. jannaschii*	31.4	1715	692	88.39	91.82	91.02	96.50
*M. genitalium*	31.7	483	301	89.44	91.52	89.89	92.32
*H. influenzae*	38.0	1754	885	91.56	96.34	93.10	98.01
*H. pylori*	38.9	1593	712	91.39	96.80	95.70	95.49
*A. aeolicus*	43.3	1517	723	95.78	96.54	95.57	87.80
*B. subtilis*	43.5	4220	1832	87.93	94.95	89.86	89.47
*Synechocystis*	47.6	3169	954	93.18	96.53	91.55	90.95
*Y. pestis*	47.6	4043	1640	91.04	94.84	93.66	88.29
*E. coli*	50.8	4290	1983	89.39	92.85	92.54	89.04
*D. radiodurans*	67.0	2622	1436	84.28	85.65	92.61	95.50
*R. solanacearum*	67.0	3442	1748	84.74	88.60	89.82	93.20
*S. coelicolor*	72.1	7851	956	88.35	91.55	91.55	90.10

The genomes are listed according to ascending G+C content. For each genome, the table shows: Genome GC content (GC %), the number of genes annotated in GenBank for that genome, the number of genes in the RescueNet training set, overall RescueNet sensitivity (Sn.), the sensitivity of RescueNet in finding genes longer than the 225 bp minimum prediction size (Sn. >225 bp), the sensitivity of RescueNet in finding genes that have been confirmed by homology with other genes in GenBank (Sn. Conserved), and finally, overall RescueNet specificity (Sp.)

ISSN: 1471-2105