Table 3.
Successful annotation based on BLAST against different gene collections
| BLAST hits | ||||
| Field | min E-valuea | Totalb | Bestc | Uniqued |
| Dmel.pro1 | 1e-171 | 1346 | 110.5 | 4 |
| Dmel.nuc1 | 1e-147 | 264 | 7 | 6 |
| Bmori.pro2 | 0e+00 | 1730 | 141.5 | 194 |
| Bmori nuc2 | 0e+00 | 1265 | 1222 | 20 |
| Bmori.wgs.nuc2 | 0e+00 | 1370 | 283 | 143 |
| lep.nuc3 | 0e+00 | 1012 | 459.5 | 147 |
| invert.pro3 | 1e-171 | 1488 | 170.5 | 4 |
| NonRed.pro4 | 1e-175 | 1487 | 29 | 3 |
| SwissP.pro4 | 1e-176 | 1136 | 25 | 1 |
| organel.nuc5 | 3e-70 | 26 | 3 | 2 |
| Rfam.nuc5 | 0e+00 | 27 | 2 | 0 |
| plant.pro | 1e-125 | 793 | 7.5 | 0 |
| Ecoli.nuc | 4e-12 | 7 | 1 | 1 |
a The threshold maximum E-value was set to 1e-05. b Total number of contigs with significant E-value. c Number of contigs having the lowest E-value for each specified BLAST field. When the lowest and second lowest E-values were the same for a particular contig it counted as 0.5 for each of the hit fields. d Number of contigs having significant BLAST hits exclusively for one of the fields. Some BLAST fields were combined to give the counts in Figure 2: 1 Dmel, 2 Bmori, 3 InvLep, 4 protein databases, 5 non-nuclear genes (as explained in the Methods). All collections used in our BLAST analysis were downloaded from public databases (see Methods) in June-August of 2005 (except for the organellar nucleotidic and the E. coli whole-genome sequences, both obtained in April 2004).