The Greedy assembly method looks for the maximum overlap between reads and extends the overlaps to form large contigs. The Greedy assembly method is recommended for Roche/454 reads or any other long reads datasets with an average read length that is greater than or equal to 70 bp.