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

## TL;DR

This paper shows how to detect mislabeled patient samples using a small set of SNPs from targeted sequencing data, preventing errors in lab results.

## Contribution

The study adapts SNP-based sample tracking for targeted NGS data and provides an open-source tool for automated error detection.

## Key findings

- A custom set of 28 SNPs was used to detect sample swaps and mislabeling in tNGS data.
- The method identified 10 mislabeled or swapped samples in a cohort of 1441 patients.
- An R function was developed and shared as a free tool for automated detection of sample handling errors.

## Abstract

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.

The online version contains supplementary material available at 10.1186/s13104-025-07348-3.