Fig. 4

Relative times for subtasks in three use cases. Relative times are listed for: i) transfer and decompression and pre-processing of data files (reads, references, annotations); ii) mapping of reads; and, iii) computing coverage/counts. All of the examples were run with 512 samples of metagenomic data [6] on C3SE cluster Glenn using 30 nodes. Use case 1 mapped metagenomic reads to assembled contigs using pBLAT. Use case 2 mapped metagenomic reads to a large reference (3.0 GiB) database of genes using GEM. Use case 3 mapped metagenomic reads to a small gene database (1.0 MiB) of amino acid sequences. The label ‘data transfer’ includes time for transfer and on-the-fly decompression of reads, references, and annotations