Figure 2

Comparison of methods for both assembly and abundance calculation using a synthetic community of 20 cloned Cpn 60 universal target sequences. Three different scenarios were investigated for the generation of a microbial profile (left-to-right): gsAssembler alone, gsAssembler plus Bowtie 2 for abundance, and Trinity plus Bowtie 2 for abundance. The number of community members recovered is shown across the top (out of 20). The major parameter affecting the accuracy of assembly is varied across the lower x-axis. For gsAssembler the minimum identity of overlaps was held constant at 90 while the minimum length parameter was varied. In the case of Trinity, the k-mer length was varied from 10 to 31 bp. The upper panel shows the percentage of reads which were untrackable (left ordinate) and the total error associated with each assembly (right ordinate). In the lower panel, microbial profiles are plotted as stacked bars with each element colored by organism according to the legend. Profiles marked as 'Target' indicate the actual composition of the amplicon library determined by Bowtie 2 mapping of all reads on to the 20 reference sequences.