Fig. 4

Percent of ORFs present in the joint reference but missing from the assembly (missing) vs percent of ORFs present in the contigs but missing from the reference set (aberrant) for the BEI mock communities (a) and the single species communities (b). An even trade off would be visualized as  a linear shift across the assemblers. A decrease in missing ORFs is usually paired with an increase in aberrant ORFs, with a larger magnitude of change in the number of aberrant ORFs. Points in the single species community cluster more closely by species (shape) than by assembler (color), indicating that the similarity in the sequences being assembled play a larger role in misassemblies than the assemblers themselves—general patterns within assemblers are still present