# PCR bias impacts microbiome ecological analyses

## TL;DR

PCR bias can distort microbiome diversity measurements, affecting how we interpret ecological differences between samples.

## Contribution

Identifies PCR bias effects on α- and β-diversity metrics and highlights robust alternatives like Aitchison distance.

## Key findings

- PCR bias significantly affects metrics like Shannon diversity and Weighted-Unifrac.
- Perturbation-invariant metrics like Aitchison distance remain unaffected by PCR bias.
- PCR bias can create or mask ecological differences depending on community composition.

## Abstract

Polymerase Chain Reaction (PCR) is a critical step in amplicon-based microbial community profiling, allowing the selective amplification of marker genes such as 16S rRNA from environmental or host-associated samples. Despite its widespread use, PCR is known to introduce amplification bias, where some DNA sequences are preferentially amplified over others due to factors such as primer-template mismatches, sequence GC content, and secondary structures. Although these biases are known to affect transcript abundance, their implications for ecological metrics remain poorly understood. In this study, we conduct a comprehensive evaluation of how PCR-bias influences both within-samples (α-diversity) and between-sample (β-diversity) analyses. We show that perturbation-invariant diversity measures remain unaffected by PCR bias, but widely used metrics such as Shannon diversity and Weighted-Unifrac are sensitive. To address this, we provide theoretical and empirical insight into how PCR-induced bias varies across ecological analyses and community structures, and we offer practical guidance on when bias-correction methods should be applied. Our findings highlight the importance of selecting appropriate diversity metrics for PCR-based microbial ecology workflows and offer guidance for improving the reliability of diversity analyses.

PCR amplification is a routine step in microbiome sequencing, but it does not treat all DNA sequences equally. Some bacterial sequences amplify more efficiently than others, creating PCR bias that distorts the relative abundances observed in sequencing data. While this bias is known to affect individual taxa, its impact on widely used ecological diversity measures has been unclear. In this study, we evaluate how PCR bias influences commonly used α-diversity (within-sample) and β-diversity (between-sample) metrics. We show that many popular measures–including Shannon diversity, Simpson’s index, Bray–Curtis dissimilarity, and Weighted UniFrac–can change substantially due to PCR bias, and that the direction and magnitude of this distortion depend on the underlying community composition. As a result, PCR bias can create or mask apparent ecological differences between groups. In contrast, we identify a class of metrics, such as Aitchison distance and differential log-ratios, that remain unaffected by PCR bias. These findings provide practical guidance on when PCR calibration is important and which diversity measures offer more robust ecological inference.