Fig. 3

Combined “MAD” assembly improves a suite of assembly statistics relative to individual assemblies. Basic assembly statistics are shown for the eukrhythmic reassemblies (per-sample) as compared to the designer metatranscriptomes. A Salmon percentage mapping distribution for the designer vs. reassembled metatranscriptomes. B Log-normalized contig length distributions compared between designer and reassembled. C Per-sequence fraction of GC-content for the designer as compared to the reassemblies. D Percentage mapping using Salmon of simulated datasets, separating by the percentage mapping of individual assemblies using each assembler tested (lower distributions), assembly groups clustered into one assembly with multiple underlying assemblers (“CAG”), and all assemblers and assemblies consolidated (“MAD”). All percentage mapping estimates were conducted independently for each set of raw reads from the underlying data. E Environmental data from Narragansett Bay using the same comparisons as Panel D. Vertical lines on panels D and E correspond to the mean values of the distribution or set of distributions represented by the color of each vertical line