Genotyping from targeted NGS data based on a small set of SNPs correctly matches patient samples

Mislabelling and swapping of laboratory samples are handling errors that can lead to erroneous interpretation of data and/or patient harm. Sequenced samples can be traced back to the respective donors by matching of single nucleotide polymorphisms (SNPs). Frameworks and software to do this have been developed for use with whole genome/exome sequencing data but not for targeted next-generation sequencing (tNGS), possibly due to the limited genomic coverage with tNGS and the need for individualization of the set of interrogated SNPs. We decided to adapt a popular tool for use with tNGS data, to demonstrate the possibility of selecting informative SNPs from a typical tNGS panel and to create an automated workflow for detection of sample handling errors.

We compiled a custom list of 28 SNPs and with its help we demonstrated the practicability of using only tNGS data to cost-effectively detect mislabelled samples. In two cohorts of totally 1441 patients with sequential samples, we could identify 3 sample swaps, 7 mislabelled samples (3 externally and 4 internally) and 1 mistake of unknown origin. We provide an R function for automated detection of sample swaps and mislabelling to the community as a free and open-source tool.