Blackbird 0.1 Manual
Blackbird is a novel integrated alignment- and local-assembly-based algorithm hat employs the barcode information encoded in Linked-reads to improve detection and placement of challenging medium-size events (50-10,000bp). Blackbird assembles the genome into segments and calls insertions and deletions in these segments. Without the need for a computationally expensive whole genome assembly, Blackbird uses a barcode-aware sliding window approach to assemble small segments of the target genome and sensitively call SVs in these segments.
Blackbird is able to work with Linked-read datasets and a combination of Linked-read and long-read datasets. We evaluated our method on both simulated and real whole genome human datasets. In a Linked-read mode Blackbird outperforms existing short-read and Linked-read methods, especially for insertions. In a hybrid-mode Blackbird demonstrated results similar to state-of-the-art long read tools, but requires less long reads to achieve same results. Therefore, our method might decrease the cost of SV calling procedure in clinical setting, without losing in the result quality.
Run
./spades_compile.sh -DSPADES_USE_JEMALLOC=OFF
Use blackbird binary from the bin folder.
Feel free to drop any inquiry to [email protected]