Measuring and mitigating PCR bias in microbiota datasets.

Area of Science:

• Microbiology
• Genetics
• Bioinformatics

Background:

• Polymerase chain reaction (PCR) amplification is crucial for high-throughput DNA sequencing of 16S ribosomal RNA (rRNA) genes in microbial community studies.
• PCR is known to introduce biases, affecting the accuracy of microbial relative abundance estimations in 16S rRNA gene sequencing studies.

Purpose of the Study:

• To develop and validate a paired modeling and experimental approach to characterize Polymerase chain reaction (PCR) non-primer-mismatch bias (NPM-bias) in microbiota surveys.
• To assess the impact of PCR NPM-bias on microbial relative abundance estimates.
• To identify methods for mitigating PCR NPM-bias in microbiome research.

Main Methods:

• Utilized mock bacterial communities for experimental validation of the modeling approach.
• Employed human gut microbiota samples to characterize PCR NPM-bias under realistic conditions.
• Developed log-ratio linear models to mitigate identified PCR NPM-bias.

Main Results:

• PCR NPM-bias was experimentally characterized and validated using mock and human gut microbial communities.
• Results indicated that PCR NPM-bias can alter microbial relative abundance estimates by a factor of four or more.
• Log-ratio linear models demonstrated effectiveness in mitigating PCR NPM-bias.

Conclusions:

• PCR amplification introduces significant non-primer-mismatch bias (NPM-bias) in 16S rRNA gene-based microbiota surveys.
• The developed modeling and experimental approach can accurately characterize and quantify PCR NPM-bias.
• Log-ratio linear modeling offers a viable strategy to correct for PCR NPM-bias, enhancing the reliability of microbiome data.