Skip to main content
Figure 4 | BMC Bioinformatics

Figure 4

From: Bison: bisulfite alignment on nodes of a cluster

Figure 4

Bison’s accuracy is consistent across read-length and its speed is generally consistent with increasing nodes. Sherman was used to generate simulated directional datasets of variable length, as described above. The accuracy of the resulting alignments is consistent regardless of read-length, A. The percentages in parentheses under each read-length label are the total percentage of aligned reads (from 20 million original reads). The speed with which bison_herd produces alignments decreases with increasing read-length, B. This rate, however, is generally consistent when increasing numbers of cluster nodes are used. Note that in the test dataset, the MPI overhead required to use 11 nodes with shorter reads was sufficient to over-tax the MPI implementation and network. When this will occur will vary by cluster.

Back to article page