Table 1.

Efficiency of the Velvet error-correction pipeline on the BAC data set

Each line in this table represents a different stage in Velvet. The initial graph was built directly from the BAC reads. The second was the result of node concatenation. The next three graphs were the result of the three consecutive steps of error correction: tip clipping, Tour Bus, and coverage cutoff. The last graph was obtained by building the graph of the reference sequence then submitting it to Tour Bus, to simulate an error-free and gap-free assembly.