4.1 Calibration and validation of the cell counting
approach
The flow cytometry calibration plots for each microorganism are given in
the supplementary material Figure S1. These results confirmed excellent
linear relationship between biomass concentration and the cell counting
results, with R2 ranging 0.979 – 0.983. Figure S1
also shows that cell counting with flow cytometry is more
reliable/consistent when cell concentrations are low. Using the
calibration relationship obtained from the single cultures, we validated
the accuracy of the flow cytometry measurements using static coculture
samples with known individual biomass concentrations. For each coculture
pair, six samples with different compositions were tested.
The individual biomass concentrations for each microorganism in the
cocultures measured using flow cytometry are plotted against the known
concentrations in Figure 2 (a) ~ (d), with the detailed
results provided in Supplementary Material Table S1. As shown in these
figures, the measured individual biomass concentrations (converted from
cell counting based on the calibration curves in Figure S1) show good
agreement with the known concentrations. However, there are relatively
large variations among the triplicates for each sample, especially for
higher concentrations, which is consistent with the similar trend
observed in the calibration curves in Figure S1. In addition, when the
same sample was measured multiple times, the measurements showed same
level of variations, suggesting the source of the variation was cell
counting. One possible reason for such large variation is the
non-uniform distribution of the cells in the liquid sample, and the
small sample volumes (25 µL) for cell counting makes such variation more
pronounced for higher concentrations, as observed in both calibration
and validation experiments. Another possible reason, which we believe is
more important for mixed culture samples, is the effect of sample
fixation process on cell counting.
In this work, we had to optimize the sample fixation protocols multiple
times in order to obtain the acceptable validation results. Figure 2 (e)
and (f) compare the cell counting results for a same static coculture
sample. Figure 2 (e) was obtained following the cell fixation protocol
initially developed for the salt water pair, while Figure 2 (f)
following the protocol optimized for the fresh water pair. The known and
measured individual biomass concentration are provided in Table 2. The
large measurement errors shown in Figure 2 (e) (-88.3% for the
methanotroph and 18.0% for the microalgae) indicate that some
methanotroph cells stuck to the microalgae cells and the flow cytometer
could not separate them properly. With the optimized protocol,
methanotroph cells were much better separated from microalgae cells,
which resulted in significantly reduced measurement error (-7.9% for
the methanotroph and -6.5% for the microalgae). Currently the flow
cytometry has been commonly used to characterize the composition of
synthetic microbiome. This example highlights the importance of
performing validation experiments to confirm the appropriateness of the
experimental protocol and the accuracy of the cell counting to avoid
misleading results.