Examining Sources of Error in PCR by Single-Molecule Sequencing

Abstract

Next-generation sequencing technology has enabled the detection of rare genetic or somatic mutations and contributed to our understanding of disease progression and evolution. However, many next-generation sequencing technologies first rely on DNA amplification, via the Polymerase Chain Reaction (PCR), as part of sample preparation workflows. Mistakes made during PCR appear in sequencing data and contribute to false mutations that can ultimately confound genetic analysis. In this report, a single-molecule sequencing assay was used to comprehensively catalog the different types of errors introduced during PCR, including polymerase misincorporation, structure-induced template-switching, PCR-mediated recombination and DNA damage. In addition to well-characterized polymerase base substitution errors, other sources of error were found to be equally prevalent. PCR-mediated recombination by Taq polymerase was observed at the single-molecule level, and surprisingly found to occur as frequently as polymerase base substitution errors, suggesting it may be an underappreciated source of error for multiplex amplification reactions. Inverted repeat structural elements in lacZ caused polymerase template-switching between the top and bottom strands during replication and the frequency of these events were measured for different polymerases. For very accurate polymerases, DNA damage introduced during temperature cycling, and not polymerase base substitution errors, appeared to be the major contributor toward mutations occurring in amplification products. In total, we analyzed PCR products at the single-molecule level and present here a more complete picture of the types of mistakes that occur during DNA amplification.

PacBio metabarcoding of Fungi and other eukaryotes: errors, biases and perspectives

“…However, these substitutions occurred at similar frequency in the ITS spacers and 5.8S rRNA gene, suggesting that natural mutations are of relatively minor importance. Our PacBio and Illumina substitution profiles match closely to the Taq polymerase error profile (Potapov & Ong, 2017), raising concern that the relatively sensitive proofreading enzyme may have been inactive during the amplification of our fungal and soil DNA samples. The choice and testing of polymerase and optimal amplification conditions are critical for the minimization of random errors in HTS data D'Amore et al, 2016;Gohl et al, 2016;Potapov & Ong, 2017).…”

Metabarcoding a diverse arthropod mock community

“…In addition to factors complicating the recovery of sequences from all species in a bulk sample, sequence variation introduced during PCR, library preparation and sequencing can make it difficult to assign sequences to their source species (Tedersoo et al, ). PCR error can be reduced by the use of high‐fidelity polymerases (Lee, Lu, Chang, Loparo, & Xie, ; Potapov et al ), but it is more difficult to escape complexities introduced by sequencing error because all second‐generation sequencers have error rates (e.g. 1%–2%) that are high enough to complicate the discrimination of closely related species.…”

“…PCR error can be reduced by the use of high-fidelity polymerases (Lee, Lu, Chang, Loparo, & Xie, 2016;Potapov et al 2017), but it is more difficult to escape complexities introduced by sequencing error because all second-generation sequencers have error rates (e.g. 1%-2%) that are high enough to complicate the discrimination of closely related species.…”

A fungal mock community control for amplicon sequencing experiments

“…However, it has become increasingly clear over time that high‐throughput amplicon sequencing techniques carry limitations and biases of their own (Amend, Seifert, & Bruns, ; Gohl et al., ; Lindahl et al., ; Tedersoo & Lindahl, ). Standards for processing and quality screening amplicon sequencing data have grown more stringent with recognition of the pernicious effects of spurious sequence variation introduced via polymerase chain reaction (PCR) errors (Quince, Lanzen, Davenport, & Turnbaugh, ), chimera formation (Edgar, Haas, Clemente, Quince, & Knight, ), miscalled bases during sequencing (Quince et al., ), contamination of samples by exogenous DNA (Salter et al., ) or DNA damage (Chen, Liu, Evans, & Ettwiller, ; Potapov & Ong, ).…”