Figure 3

Performance of four short-read assemblers on both single- and paired-end metatranscriptomic datasets. Assembly performance was assessed using both single-end and paired-end datasets generated from the NOD503CecMN sample. Comparisons between the two datasets are presented for each assembler/parameter combination except IDBA-MT which requires paired-end data. Assemblers were evaluated on the basis of: 1) number of contigs assembled, 2) percentage of reads that map to assembled contigs, and 3) whether contigs have sequence similarity to a known protein at two levels of stringency.