Chemical Links Between Redox Conditions and Estimated Community Proteomes from 16S rRNA and Reference Protein Sequences
Curation statements for this article:-
Curated by eLife
Evaluation Summary:
This study presents an in-depth analyses of carbon oxidation state and hydration state of proteomes in different taxa and environmental settings, which contributes to our understanding of how microbial communities are shaped by their surroundings. The study has merit, but there also some technical weaknesses.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (eLife)
Abstract
Article activity feed
-
Evaluation Summary:
This study presents an in-depth analyses of carbon oxidation state and hydration state of proteomes in different taxa and environmental settings, which contributes to our understanding of how microbial communities are shaped by their surroundings. The study has merit, but there also some technical weaknesses.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)
-
Reviewer #1 (Public Review):
In this paper the authors present variations in carbon oxidation state and hydration state in proteomes available in RefSeq. Then they use this information to predict community level proteomes, and their corresponding carbon oxidation states and hydration states, based on available 16S rRNA gene sequences from selected previously published datasets. When combining this with information about the environmental setting of the individual samples analyzed, the authors are able to demonstrate connections between redox conditions and proteomic carbon oxidation state and hydration state. Furthermore, they explore how individual taxonomic groups at different taxonomic levels contribute to forming these connections.
A weakness with the study is that the described environmental proteomes are inferred from 16S rRNA …
Reviewer #1 (Public Review):
In this paper the authors present variations in carbon oxidation state and hydration state in proteomes available in RefSeq. Then they use this information to predict community level proteomes, and their corresponding carbon oxidation states and hydration states, based on available 16S rRNA gene sequences from selected previously published datasets. When combining this with information about the environmental setting of the individual samples analyzed, the authors are able to demonstrate connections between redox conditions and proteomic carbon oxidation state and hydration state. Furthermore, they explore how individual taxonomic groups at different taxonomic levels contribute to forming these connections.
A weakness with the study is that the described environmental proteomes are inferred from 16S rRNA gene sequence data and not observed directly. However, there is good reason to believe that the conclusions drawn in the paper are valid.
The study sheds light on microbial adaptations on the genome level that so far have received relatively little attention. The paper is also interesting from an ecological perspective regarding the general question of how microbial communities are shaped by environmental settings.
-
Reviewer #2 (Public Review):
This manuscript mainly investigated the carbon oxidation and stoichiometric hydration states of the inferred community proteomes according to 16S rRNA gene compositions from the published datasets and explored their potential associations with environmental parameters such as redox gradients, oxygen concentrations and salinity.
Predictions of the carbon oxidation and stoichiometric hydration states on the basis of microbial proteomes can provide some meaningful information for disentangling microbial response to environmental changes. As we know, some genes in microbial genomes are not expressed and transformed to proteins. Therefore, such gene redundancy in genomes may lead to bias in predicting the carbon oxidation and stoichiometric hydration states. Furthermore, this study compiled many 16S rRNA gene …
Reviewer #2 (Public Review):
This manuscript mainly investigated the carbon oxidation and stoichiometric hydration states of the inferred community proteomes according to 16S rRNA gene compositions from the published datasets and explored their potential associations with environmental parameters such as redox gradients, oxygen concentrations and salinity.
Predictions of the carbon oxidation and stoichiometric hydration states on the basis of microbial proteomes can provide some meaningful information for disentangling microbial response to environmental changes. As we know, some genes in microbial genomes are not expressed and transformed to proteins. Therefore, such gene redundancy in genomes may lead to bias in predicting the carbon oxidation and stoichiometric hydration states. Furthermore, this study compiled many 16S rRNA gene datasets from previous studies. Different primer sets were applied in those studies, and such difference will result in distinct 16S rRNA gene compositions. Accordingly, it is essential to deal with the influence of different primer sets on the 16S rRNA gene compositions among samples. Unfortunately, such information is missing in the method section. Additionally, the community proteomes in this study were inferred from 16S rRNA genes. The marker gene of 16S rRNA cannot well predict their corresponding genomes, possibly leading to prediction of biased proteomes. Therefore, it should avoid to use 16S rRNA genes for predicting microbial genomes and proteomes.
It seems that the relationships between carbon oxidation states/stoichiometric hydration state and redox/salinity gradients have been reported in previous studies (e.g., Dick et al 2019, 2020, 2021). The finding of this study is not new in comparison with the previously reported.
-
Author Response
Reviewer #1 (Public Review):
In this paper the authors present variations in carbon oxidation state and hydration state in proteomes available in RefSeq. Then they use this information to predict community level proteomes, and their corresponding carbon oxidation states and hydration states, based on available 16S rRNA gene sequences from selected previously published datasets. When combining this with information about the environmental setting of the individual samples analyzed, the authors are able to demonstrate connections between redox conditions and proteomic carbon oxidation state and hydration state. Furthermore, they explore how individual taxonomic groups at different taxonomic levels contribute to forming these connections.
A weakness with the study is that the described environmental proteomes are inferred …
Author Response
Reviewer #1 (Public Review):
In this paper the authors present variations in carbon oxidation state and hydration state in proteomes available in RefSeq. Then they use this information to predict community level proteomes, and their corresponding carbon oxidation states and hydration states, based on available 16S rRNA gene sequences from selected previously published datasets. When combining this with information about the environmental setting of the individual samples analyzed, the authors are able to demonstrate connections between redox conditions and proteomic carbon oxidation state and hydration state. Furthermore, they explore how individual taxonomic groups at different taxonomic levels contribute to forming these connections.
A weakness with the study is that the described environmental proteomes are inferred from 16S rRNA gene sequence data and not observed directly. However, there is good reason to believe that the conclusions drawn in the paper are valid.
The study sheds light on microbial adaptations on the genome level that so far have received relatively little attention. The paper is also interesting from an ecological perspective regarding the general question of how microbial communities are shaped by environmental settings.
To attempt to bring more attention to environmental constraints, a plot (Figure 4E in the published paper) was redrawn to more clearly show how carbon oxidation state of estimated community proteomes not only is lower in more reducing conditions for a variety of environments but also shows the largest differences for hydrothermal systems and shale-gas wells. This finding is discussed in terms of geological sources of reductants and provides new evidence that the chemical makeup of microbial communities reflects their geological context.
Reviewer #2 (Public Review):
This manuscript mainly investigated the carbon oxidation and stoichiometric hydration states of the inferred community proteomes according to 16S rRNA gene compositions from the published datasets and explored their potential associations with environmental parameters such as redox gradients, oxygen concentrations and salinity.
Predictions of the carbon oxidation and stoichiometric hydration states on the basis of microbial proteomes can provide some meaningful information for disentangling microbial response to environmental changes. As we know, some genes in microbial genomes are not expressed and transformed to proteins. Therefore, such gene redundancy in genomes may lead to bias in predicting the carbon oxidation and stoichiometric hydration states.
Our study uses available data sources to identify informative differences of elemental compositions of proteomes predicted from genomes. There are numerous examples in the literature of using protein sequences predicted from genomes to make comparisons of amino acid composition (for example, in eLife: https://doi.org/10.7554/eLife.57347), so it would appear to be acceptable with some level of uncertainty to use genomic data to make comparisons between (amino acid or elemental) compositions of predicted proteomes.
Furthermore, this study compiled many 16S rRNA gene datasets from previous studies. Different primer sets were applied in those studies, and such difference will result in distinct 16S rRNA gene compositions. Accordingly, it is essential to deal with the influence of different primer sets on the 16S rRNA gene compositions among samples. Unfortunately, such information is missing in the method section.
Primer sets used in the source studies have been added to Table 1 in the published paper. The Discussion was modified to acknowledge limitations in making comparisons *between* datasets obtained using different primers. However, the main results of this study are based on differences of carbon oxidation state (Zc) *within* individual datasets (for instance, along the vertical redox gradients shown in Figure 3).
The intra-dataset differences of Zc themselves are compared across datasets in Figure 4E. However, it can be expected that the effects of technical variability – including not only primer pairs but also DNA extraction methods, etc. – would tend to be reduced in these inter-dataset comparisons of intra-dataset differences, in contrast to direct inter-dataset comparisons. The index plot at the center of Figure 2 does make a direct inter-dataset comparison, but the outcome is consistent with trends identified in previous analyses of shotgun metagenomic datasets, 16S primers and other technical differences between studies notwithstanding.
Additionally, the community proteomes in this study were inferred from 16S rRNA genes. The marker gene of 16S rRNA cannot well predict their corresponding genomes, possibly leading to prediction of biased proteomes. Therefore, it should avoid to use 16S rRNA genes for predicting microbial genomes and proteomes.
Despite the various sources of uncertainty in making estimates of elemental composition of communities from 16S rRNA genes and reference proteomes, comparisons with shotgun metagenomic data support the reliable identification of trends within datasets (Figure 5 in the published paper).
It seems that the relationships between carbon oxidation states/stoichiometric hydration state and redox/salinity gradients have been reported in previous studies (e.g., Dick et al 2019, 2020, 2021). The finding of this study is not new in comparison with the previously reported.
The explorations in previous studies of chemical links between communities and environments were based on analysis of shotgun metagenomic data. The ability to reproduce those findings by analyzing 16S rRNA gene sequence data is a new advance in this study.
Other new results in the published paper are the different magnitudes of Zc differences in various environments (which were not previously documented from shotgun metagenomes; Figure 4E) and the comparison of shotgun metagenome and 16S-based estimates of Zc for the time series of injected fluids in the Marcellus Shale (Figure 5B). The latter results are particularly interesting; the close correspondence for Days 0, 7, and 13 supports the basic reliability of the 16S-based estimates, while the increasing divergence at Days 82 and 328 suggests the onset of some interfering mechanisms (the speculation is made that this could be related to viral lysis and heterotrophic degradation of the released DNA). Also, the published paper presents the first analysis of carbon oxidation state of proteins – from either shotgun metagenome sequences or 16S rRNA-based estimates – for microbial communities in various body sites using data from the Human Microbiome Project (Figure 5D).
-
-
-