Abstract
Quantitative stable isotope probing (qSIP) estimates isotope tracer incorporation into DNA of individual microbes and can link microbial biodiversity and biogeochemistry in complex communities. As with any quantitative estimation technique, qSIP involves measurement error, and a fuller understanding of error, precision, and statistical power benefits qSIP experimental design and data interpretation. We used several qSIP data sets-from soil and seawater microbiomes-to evaluate how variance in isotope incorporation estimates depends on organism abundance and resolution of the density fractionation scheme. We assessed statistical power for replicated qSIP studies, plus sensitivity and specificity for unreplicated designs. As a taxon's abundance increases, the variance of its weighted mean density declines. Nine fractions appear to be a reasonable trade-off between cost and precision for most qSIP applications. Increasing the number of density fractions beyond that reduces variance, although the magnitude of this benefit declines with additional fractions. Our analysis suggests that, if a taxon has an isotope enrichment of 10 atom% excess, there is a 60% chance that this will be detected as significantly different from zero (with alpha 0.1). With five replicates, isotope enrichment of 5 atom% could be detected with power (0.6) and alpha (0.1). Finally, we illustrate the importance of internal standards, which can help to calibrate per sample conversions of %GC to mean weighted density. These results should benefit researchers designing future SIP experiments and provide a useful reference for metagenomic SIP applications where both financial and computational limitations constrain experimental scope.IMPORTANCE One of the biggest challenges in microbial ecology is correlating the identity of microorganisms with the roles they fulfill in natural environmental systems. Studies of microbes in pure culture reveal much about their genomic content and potential functions but may not reflect an organism's activity within its natural community. Culture-independent studies supply a community-wide view of composition and function in the context of community interactions but often fail to link the two. Quantitative stable isotope probing (qSIP) is a method that can link the identity and functional activity of specific microbes within a naturally occurring community. Here, we explore how the resolution of density gradient fractionation affects the error and precision of qSIP results, how they may be improved via additional experimental replication, and discuss cost-benefit balanced scenarios for SIP experimental design.