Fig. 5

The read forest construction process of Assembltrie. To place each read in a cycle-rooted trie, Assembltrie greedily identifies its parent and children by the use of prefix and suffix K-mer hash tables. The initial K-mer matches (i.e., hash table hits) are extended allowing a maximum number \({\it{\epsilon }}\) of mismatches. As a user option, the greedy strand synchronization heuristic picks for each read, the strand that has the longest prefix–suffix overlap with the implied parent