Iterative low-memory k-mer trimming.

The results of trimming reads at unique (erroneous) k-mers from a 5 m read E. coli data set (1.4 GB) in under 30 MB of RAM. After each iteration, we measured the total number of distinct k-mers in the data set, the total number of unique (and likely erroneous) k-mers remaining, and the number of unique k-mers present at the 3' end of reads.