Unique, dual-matched adapters mitigate index hopping between NGS samples

The increasing throughput of NGS platforms has fueled the demand to sequence many samples in parallel, also referred to as multiplex sequencing. During multiplex sequencing, the identity of each sample library within a pool is maintained using index sequences that are subsequently separated in a process called demultiplexing during data analysis. Historically, a relatively small number of unique sequences (8 x i5 and 12 x i7) were used to create index combinations to multiplex samples. Unfortunately, with this combinatorial approach, a single index swap may cause a read to be mis-assigned to a different sample causing cross-talk. In this presentation, we discuss some sources of sample cross-talk, including index hopping during cluster amplification or multiplexed capture, and how index sequencing errors may lead to demultiplexing mistakes. We discuss how sample cross-talk causes demultiplexing errors and present a method for increasing the accuracy of sample identification using unique, dual-matched index adapters.