# PCR-free library preparation greatly reduces stutter noise at short tandem repeats

## TL;DR

A PCR-free library preparation method significantly reduces stutter noise in STR genotyping, improving accuracy for forensic and genetic studies.

## Contribution

The study demonstrates that PCR-free protocols reduce stutter artifacts, particularly at homopolymers, enabling more reliable STR genotyping.

## Key findings

- PCR-free protocol reduces stutter reads by over four-fold compared to traditional methods.
- Stutter at homopolymers is reduced by more than six-fold, allowing better STR calling.
- High confidence STR genotypes are achievable with PCR-free sequencing.

## Abstract

Over the past several decades, the forensic and population genetic communities have increasingly leveraged short tandem repeats (STRs) for a variety of applications. The advent of next-generation sequencing technologies and STR-specific bioninformatic tools has enabled the profiling of hundreds of thousands of STRs across the genome. Nonetheless, these genotypes remain error-prone, hindering their utility in downstream analyses. One of the primary drivers of STR genotyping errors are “stutter” artifacts arising during the PCR amplification step of library preparation that add or delete copies of the repeat unit in observed sequencing reads. Recently, Illumina developed the TruSeq PCR-free library preparation protocol which eliminates the PCR step and theoretically should reduce stutter error. Here, I compare two high coverage whole genome sequencing datasets prepared with and without the PCR-free protocol. I find that this protocol reduces the percent of reads due to stutter by more than four-fold and results in higher confidence STR genotypes. Notably, stutter at homopolymers was decreased by more than 6fold, making these previously inaccessible loci amenable to STR calling. This technological improvement shows good promise for significantly increasing the feasibility of obtaining high quality STR genotypes from next-generation sequencing technologies.