Error Correction Using Short Reads

A number of strategies have been developed to combine the beneficial features of Illumina and PacBio reads during the genome assembly process. Illumina reads are short but have low error rate, whereas PacBio reads are long and have 15% rate of indel errors.

PacbioToCA and LSC use an alignment step to align all Illumina reads with PacBio reads, and then use the alignment to correct PacBio reads. The alignment step is extremely slow, because a large number of Illuminia reads are aligned using a gapped aligner, such as Novoalign in case of LSC.

It is far more efficient to assemble the Illumina reads into contigs, and then use BLASR to align PacBio reads on to those contigs (C. S. Chin, personal communication).