Relative stabilities of proteins inferred from a hot-spring metagenome (2011, 2013)

Measured temperature and pH

bison1()

Link to original figure. The lines are spline fits that represent smooth changes in T and pH that are used for some plots below.

Carbon oxidation state of proteins

bison2()

Links to original figures: functional annotations and major phyla.

Chemical affinities

tab <- round(bison3(), 3)
htmltab <- capture.output(kable(tab, format = "html"))
# make selected values bold
for(i in c(15, 16, 41, 42, 43)) {
  htmltab[i] <- gsub('">', '"><b>', htmltab[i], fixed = TRUE)
  htmltab[i] <- gsub('</', '</b></', htmltab[i], fixed = TRUE)
}
structure(htmltab, format = "html", class = "knitr_kable")

Link to original table. These per-residue affinities are expressed as dimensionless values, i.e. A/(2.303RT) where A, R and T stand for chemical affinity, gas constant, and temperature in Kelvin. The affinities are all negative, but are higher (more positive) for the reaction to form model protein 1 at sites 1 and 2 and to form model protein 4 at sites 3–5.

Relative stabilities along a temperature and chemical gradient

bison4()

Link to original figure. The first plot is a theoretical predominance diagram for the overall model proteins from each site as a function of temperature and activity of hydrogen. The dotted line is placed so that it passes through the stability fields of the proteins at approximately the actual temperatures where they are located in the hot spring. This is Equation 2 in the 2011 paper: loga_H2 = -11 + (3/40) T,°C. The second plot shows the effect of changing both temperature and pH (interpolated from measured values) together with loga_H2 (from the line in the first plot) on the predicted relative abundances of the proteins in metastable equilibrium.

Comparing old and new group additivity parameters

bison5()

Link to original figure. The first row of plots is made using group additivity parameters from Dick et al. (2006); the second is made with updates for the methionine sidechain from LaRowe and Dick (2012) and the glycine sidechain and protein backbone group from Kitadai (2014). Note the appearance of a stability field for site 3 for some classes of proteins. The remaining figures in this vignette use the updated parameters (i.e. the current default values in OBIGT).

Metastable equilibrium model for relative abundances

equil.results <- bison6()

Link to original figure. The top row shows calculated relative abundances of the phyla based on metastable equilibrium among the model proteins as a function of loga_H2. The bottom row shows the Gibbs energy of transformation (ΔG_tr) between the equilibrium and actual relative abundances. Lines are drawn at values calculated with Equation 2 (see above) and at the minimum ΔG_tr, i.e. the optimal values of loga_H2 for the model. Plots are made for only three sites, but the function returns results for all five sites, which are used below.

Activity of hydrogen comparison

bison7(equil.results)

Links to original figures: 2011, 2013 The plot on the right, from the 2011 paper, shows effective values of loga_H2 calculated from oxidation-reduction potential (ORP) measurements in 2009 and sulfate/sulfide ratios and dissolved oxygen concentrations measured in 2005. The different redox reactions are clearly not in mutual equilibrium, but they all indicate relatively oxidizing conditions at lower temperatures. Equation 2 is derived from the relative stabilities of proteins across sites (see above). The plot on the right, from the 2013 paper, shows the results of optimizing the metastable equilibrium model for phylum abundances within sites.

Comparison of model and observed abundances

bison8(equil.results)

Link to original figure.

Document history

• 2011-08-23 Initial version.
• 2012-01-07 Use superseded properties of methionine sidechain group.
• 2014-02-01 Include results from 2013 paper and change VignetteEngine to knitr.
• 2020-07-12 Move from CHNOSZ to JMDplots; convert document from LaTeX (Rnw) to R Markdown (Rmd) and add colors.