Improved Sequencing of Low Diversity Samples

摘要

Low cost and quick turnaround make the Illumina MiSeq an attractive system for many small-scale sequencing projects. However, low-complexity samples can pose several challenges for the Illumina platform. Specifically, cluster identification, and phasing and color matrix calculations can be negatively impacted by a lack of sequence diversity. These issues can be mitigated by reducing cluster density and by including a high-diversity library such as a PhiX control library, but this reduces the number of useful reads. PhiX levels of 40% or higher may be needed to ensure a successful run. We have optimized an alternative approach that minimizes the effect of low diversity samples while retaining a high number of useful reads. This is accomplished using adaptors that contain additional sequences that are randomized for both sequence and length. This allows for sequencing of very low diversity samples at high cluster density without the need for a PhiX library spike-in. Examples of applications that can benefit from such an approach include deep amplicon sequencing for the detection of rare mutations, 16S ribosomal sequencing for bacterial population analysis, and reduced representation bisulfite sequencing (RRBS). RRBS is a particularly interesting example of low diversity, where all forward reads start with either TGG or CGG. The complete lack of color balance at bases 2 and 3 makes these samples very challenging for both MiSeq and HiSeq instruments. The approach described here of adding complex, variable length sequences at the start of the read allows RRBS libraries to be sequenced more efficiently.