Interpreting ecological diversity indices applied to terminal restriction fragment length polymorphism data: insights from simulated microbial communities

Abstract

Ecological diversity indices are frequently applied to molecular profiling methods, such as terminal restriction fragment length polymorphism (T-RFLP), in order to compare diversity among microbial communities. We performed simulations to determine whether diversity indices calculated from T-RFLP profiles could reflect the true diversity of the underlying communities despite potential analytical artifacts. These include multiple taxa generating the same terminal restriction fragment (TRF) and rare TRFs being excluded by a relative abundance (fluorescence) threshold. True community diversity was simulated using the lognormal species abundance distribution. Simulated T-RFLP profiles were generated by assigning each species a TRF size based on an empirical or modeled TRF size distribution. With a typical threshold (1%), the only consistently useful relationship was between Smith and Wilson evenness applied to T-RFLP data (TRF-E(var)) and true Shannon diversity (H'), with correlations between 0.71 and 0.81. TRF-H' and true H' were well correlated in the simulations using the lowest number of species, but this correlation declined substantially in simulations using greater numbers of species, to the point where TRF-H' cannot be considered a useful statistic. The relationships between TRF diversity indices and true indices were sensitive to the relative abundance threshold, with greatly improved correlations observed using a 0.1% threshold, which was investigated for comparative purposes but is not possible to consistently achieve with current technology. In general, the use of diversity indices on T-RFLP data provides inaccurate estimates of true diversity in microbial communities (with the possible exception of TRF-E(var)). We suggest that, where significant differences in T-RFLP diversity indices were found in previous work, these should be reinterpreted as a reflection of differences in community composition rather than a true difference in community diversity.

FIG. 1.

Illustration of how the species abundance distribution and TRF distribution interact to form a simulated T-RFLP profile. Illustrated steps include calculating relative abundances in a lognormal species abundance distribution based on systematically iterated distribution parameters (1), defining a TRF size probability distribution (one of the four empirical restriction enzyme databases was used for OTU sampling and sequence sampling simulations, whereas a new TRF size distribution was generated for each iteration in parametric sampling simulations) (2), randomly assigning each species in the simulated community to a TRF size according to the TRF size probability distribution (3), summing signals from TRFs and applying a relative abundance (fluorescence) threshold and upper and lower fragment size cutoffs (4), and comparing diversity statistics calculated on the underlying community (species abundance distribution) and the simulated T-RFLP profile by calculating correlation coefficients and plotting data density graphs (5).

FIG. 2.

Data density plots for OTU sampling simulations of communities with up to 1,200 species, using a relative abundance threshold of 1% and the HhaI TRF distribution. These plots summarize a data cloud. Lines can be interpreted as contour lines of data point density around a “peak” showing the central relationship between the two variables. Overlapping percentile ranges for two values of the TRF index indicate the potential for the true indices to be the same for the two communities. Therefore, the slope of the central relationship, as well as the width of the percentile ranges, is important. The TRF diversity index data were broken up into five percentile groups. The mean value for the TRF index in each group was plotted against the following true community index values: mean (diamond), 25th and 75th percentiles (bold line), 10th and 90th percentiles (thin line), and 2.5th and 97.5th percentiles (dashed line). In panels A and B, n was 940 for each group. In panel C, n was 638 to 1,998 due to TRF-S ties.

FIG. 3.

Data density plots for parametric sampling simulations of communities with up to 10,000 species, using a relative abundance threshold of 1%. See the legend to Fig. 2 for explanation. n was 7,973 for each group.