Page 122 - Read Online

Wang et al. Microbiome Res Rep. 2025;4:23 https://dx.doi.org/10.20517/mrr.2024.94 Page 7 of 19

Figure 2. Correlation analysis of read counts of the two spike-in bacteria and the microbial loads from spike-in. The correlation analysis
was performed using Pearson’s correlation coefficient. (A) The line chart illustrates the read counts of Planococcus as a function of
microbial load from spike-ins across different samples; (B) The line chart illustrates the read counts of Pseudoalteromonas as a function of
microbial load from spike-ins across different samples.

-0.78 (P = 0.002781). The close similarity between the two r values suggests that the spike-in method results
in a consistent relationship with microbial load across different spike-in bacterial strains, although they are
of different gram types. While the correlation for Planococcus was found to be slightly stronger than that for
Pseudoalteromonas, both values confirm the effectiveness of the spike-in method for accurately quantifying
microbial load.

Spike-in comparison with other absolute quantification approaches (Flow Cytometry, Plate Count,
qPCR, and Total DNA)
As shown in Figure 1A, flow cytometry produced significantly higher microbial cell counts compared to
plate count (P = 0.01), spike-in (P = 0.002), and qPCR (P = 0.002). Plate count also yielded significantly
higher counts than spike-in (P = 0.016) and qPCR (P = 0.036). No statistically significant difference was
observed between spike-in and qPCR (P = 0.07), indicating general agreement between these two methods.
Overall, the methods followed a descending trend in estimated bacterial load: Flow Cytometry > Plate
Count > qPCR ≈ Spike-in.

When mother and infant groups were analyzed separately, significant differences in total microbial cell
counts were detected using qPCR (P < 0.001) and spike-in quantification (P = 0.005), but not with flow
cytometry or plate count. This highlights the higher sensitivity of DNA-based methods in detecting group-
specific microbial load differences and further underscores the consistency between qPCR and spike-in
results.

A strong positive correlation was observed between qPCR and spike-in results across individual samples,
with Pearson's correlation coefficient calculated as r = 0.97 (P < 0.0001) [Figure 3B]. This indicates a high
degree of agreement between these two methods. Furthermore, total microbial load was significantly greater
in the mother group compared to the infant group (P < 0.001), consistent across both qPCR and spike-in
analyses.