Fig. 1

Reference-guided de novo assembly pipeline. Raw reads get quality trimmed (1. step) and mapped against a reference (2. step). Reference mapped reads are grouped into blocks with continuous read coverage. These blocks are then combined into superblocks until a total length of at least 12 kb is reached. Superblocks are overlapping by at least one block. Each superblock and all unmapped reads are separately de novo assembled (3. step). Resulting contigs are merged into non-redundant supercontigs (4. step). In the fifth step, reads are mapped back to the supercontigs and unmapped reads are de novo assembled to get additional supercontigs. All supercontigs are error corrected with back mapped reads (6. step) and afterwards used for scaffolding and gap closing (7. step)