Table 2.
Sequencing error rates for the ChIP and RNA experiments
| Read mapping method | ChIP error rate | RNA error rate |
|---|---|---|
| (Algorithms) | (% data used) | (% data used) |
| Highest intensity base | — | 2.26 |
| (None) | (100) | |
| Most probable base (MPB) | 2.29 | 4.17 |
| (SeqMap, SOAP, MAQ, Bowtie) | (100) | (100) |
| MPB + chastity filtering | — | 4.45 |
| (Solexa Pipeline) | (97.8) | |
| MPB + quality filtering | 0.83 | 2.44 |
| (RMAP) | (69.8) | (64.2) |
| Probability mapping | 0.67 | 1.64 |
| (GNUMAP, Novo, Slider) | (100) | (100) |
| Probability + quality | 0.27 | 1.11 |
| (GNUMAP) | (69.8) | (64.2) |
Notice that the probabilistic mapping (GNUMAP) error rates are lower than those for other methods. In particular, the error rates for GNUMAP (using all the data) are lower using than those that incorporate base-calling and filtering (deleting 30–35% of the data).