A novel post hoc method for detecting index switching finds no evidence for increased switching on the Illumina HiSeq X

Abstract

High-throughput sequencing using the Illumina HiSeq platform is a pervasive and critical molecular ecology resource, and has provided the data underlying many recent advances. A recent study has suggested that “index switching,” where reads are misattributed to the wrong sample, may be higher in new versions of the HiSeq platform. This has the potential to invalidate both published and in-progress work across the field. Here, we test for evidence of index switching in an exemplar whole-genome shotgun data set sequenced on both the Illumina HiSeq 2500, which should not have the problem, and the Illumina HiSeq X, which may. We leverage unbalanced heterozygotes, which may be produced by index switching, and ask whether the undersequenced allele is more likely to be found in other samples in the same lane than expected based on the allele frequency. Although we validate the sensitivity of this method using simulations, we find that neither the HiSeq 2500 nor the HiSeq X has evidence of index switching. This suggests that, thankfully, index switching may not be a ubiquitous problem in HiSeq X sequence data. Lastly, we provide scripts for applying our method so that index switching can be tested for in other data sets.