genomewide characterization of host-pathogen interactions by ...

log 2 ratio (inf. RN1HG ΔsigB / sham inf.)
Maren Depke
Results
Kidney Gene Expression Pattern in an in vivo Infection Model
Comparative analysis of S. aureus infected samples with sham infection identifies strong
infection/inflammation reactions of the kidney tissue
As high reproducibility of kidney expression profiles in biological replicates of infection with
S. aureus had been shown, the comparison between S. aureus infected samples and sham
infection / NaCl control was conducted with only one representative biological replicate (for
comparison see Fig. R.2.3; and Fig. R.2.4, panel ).
When comparing sham infection with infection with S. aureus RN1HG, 5264 sequences were
differentially expressed which correspond to 4613 genes (EntrezGene records in Rosetta Resolver
software). Of these genes, 1083 possessed an absolute fold change of at least 2. A similar
difference was observed in the comparison of sham infection with infection with S. aureus
RN1HG ΔsigB: 4826 sequences (4248 genes) differed significantly of which 970 genes exhibited
an absolute fold difference of at least 2 (Table R.2.1).
When calculating ratios of highly similar sample data sets to a common baseline, a result of
conforming values is expected. The log-ratio plot of the two ratios “infection with
RN1HG” / “sham infection” and “infection with RN1HG ΔsigB” / “sham infection” in the second
biological replicate BR2 on EntrezGene level visualizes this similarity of the ratio data: Data points
are mainly arranged near the diagonal of the plot (Fig. R.2.6). Some data points display more
scattering, but the difference in the ratios was not significant when tested statistically and was
mainly due to higher variation in signal intensities of one group, as described above.
6
Fig. R.2.6:
Ratio plot comparing the log 2ratio of “infection with
RN1HG” / “sham infection” with the log 2ratio of
“infection with RN1HG ΔsigB” / “sham infection” in the
second biological replicate BR2 on EntrezGene level.
Genes differentially regulated (p* < 0.01 in errorweighted
one-way ANOVA with Benjamini-Hochberg
False Discovery Rate multiple testing correction in the
Rosetta Resolver software) in both comparisons are
colored in light gray, genes specifically significant in the
comparison “infection with RN1HG” / “sham infection”
are depicted in black while genes specifically significant
in the comparison “infection with RN1HG
ΔsigB” / “sham infection” are shown in dark gray. In
total, 5025 genes are visible.
BR – biological replicate.
4
2
0
-2
-4
-6
-6 -4 -2 0 2 4 6
log 2 ratio (inf. RN1HG / sham inf.)
The genes differentially regulated in S. aureus infected kidney tissue compared to sham
infection were further analyzed using the Ingenuity Pathway Analysis tool. Only genes displaying
an at least twofold difference in expression between sham infected and S. aureus infected
animals (either RN1HG or RN1HG ΔsigB from the second biological replicate BR2) were
considered in this analysis. Using this approach, the list of genes eligible for analysis was
restricted to 1156 genes of which 4 could not be mapped to internal data base entries (Ingenuity
Pathway Knowledge Base, IPKB) by IPA. Since expression values of experiments involving
infection with S. aureus RN1HG or its isogenic sigB mutant did not differ significantly, the average
of both expression values was compared to sham infected animals, which allows intuitive
visualization of the data.
The global functional analysis using IPA offers an overview on the biological functions that are
associated with the analyzed data set. As expected from the experimental setting, influence on
the functional category “Inflammatory Response” was most pronounced with p-values of
2.36E−75 to 1.24E−10 for the sub-categories and with 343 associated genes differentially
expressed between infection and sham infection. Additionally, other infection, immune response
82