Exploring Bacterial Strain Diversity in the Asymptomatic Urinary Microbiome
This article has been Reviewed by the following groups
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Taxa of the asymptomatic (“healthy”) urinary tract microbiota (urobiome) have been identified through high-throughput, short-read sequencing of the 16S rRNA and culture-based studies. These methods have identified numerous genera and species, respectively, that persist within this low biomass environment. Studies employing shotgun metagenomic sequencing of urine samples has only recently been done, and they have primarily focused on identifying non-prokaryotic constituents and/or assessing the functional capacity of the community. Species diversity within the urobiome has yet to be explored. Here we conducted a proof-of-concept study using the software tool STRONG to identify taxa and strain diversity within the asymptomatic urobiome. Two mock communities of urinary strains were produced and sequenced to provide a benchmark for such analyses. Sixty-one urobiome data sets were examined. Species were identified in 29 (47.54%) of the samples examined, including 9 samples containing more than one species, and 8 samples containing more than one strain of a given species. This suggests that the asymptomatic urobiome can include strain diversity, despite being a low biomass community. Further investigation into strain diversity within the urobiome will provide insight into dynamics of this community, in particular with respect to lower urinary tract symptoms.
Article activity feed
-
The reviewers have highlighted major concerns with the work presented. Please ensure that you address their comments. The reviewers raise concerns regarding the scientific rigour and experimental design of the work.
-
Comments to Author
The manuscript authored by Shanbhag et al. presents a study that addresses the important need to explore strain diversity in urobiome research. The authors propose employing shotgun metagenomic sequencing of urine samples to achieve this goal. Additionally, the study includes the preparation of two mock communities to control the metagenomic sequencing process. While the idea of investigating strain diversity is pertinent, there are several concerns regarding the study design, chosen methodology, and data analysis.One major concern revolves around the concept of urine being a low biomass sample. While it is true that urine harbors fewer bacteria compared to the gut, studies have shown high bacterial counts when culture conditions are optimized for the bacterial species present in the urinary tract. …
Comments to Author
The manuscript authored by Shanbhag et al. presents a study that addresses the important need to explore strain diversity in urobiome research. The authors propose employing shotgun metagenomic sequencing of urine samples to achieve this goal. Additionally, the study includes the preparation of two mock communities to control the metagenomic sequencing process. While the idea of investigating strain diversity is pertinent, there are several concerns regarding the study design, chosen methodology, and data analysis.One major concern revolves around the concept of urine being a low biomass sample. While it is true that urine harbors fewer bacteria compared to the gut, studies have shown high bacterial counts when culture conditions are optimized for the bacterial species present in the urinary tract. Thus, the issue lies not in the number of bacteria but rather in the growth conditions used for cultivation. Consequently, this raises questions about the composition of the mock communities used in this study. The species chosen are not representative of those commonly found in healthy urobiomes. This deviation from natural populations hampers the ability to draw meaningful conclusions from the mock communities' results. Moreover, the selected species fail to mimic the challenges associated with the bacteria present in the healthy urobiome, such as difficulties in DNA extraction. Therefore, the utility of these mock communities in providing accurate controls for metagenomic sequencing is questionable. In addition, the manuscript lacks comprehensive exploration of the sequencing data. Crucial information regarding the characteristics that differentiate identified strains and their clinical impact, such as virulence factors and antimicrobial resistance genes, is missing. Understanding these attributes is vital for interpreting the functional potential and health implications of strain diversity within the urobiome. The absence of such analyses weakens the study's significance and hinders its potential contributions to the field.
Please rate the manuscript for methodological rigour
Poor
Please rate the quality of the presentation and structure of the manuscript
Satisfactory
To what extent are the conclusions supported by the data?
Partially support
Do you have any concerns of possible image manipulation, plagiarism or any other unethical practices?
No
Is there a potential financial or other conflict of interest between yourself and the author(s)?
No
If this manuscript involves human and/or animal work, have the subjects been treated in an ethical manner and the authors complied with the appropriate guidelines?
Yes
-
Comments to Author
Here the authors report the results of implementation of a relatively new microbiome analysis tool, STRONG, to resolve strain-level features in two urinary microbiome datasets. As the field moves more towards species-level analysis through implementation of shotgun metagenomic sequencing and improved amplicon sequencing methods, the resolution of bacterial strains within the urinary microbiome has become of greater interest. The implementation of methodologies capable of identifying multiple strains of the sample species within shotgun metagenomic sequencing datasets on less understood microbiomes like the urinary microbiome is important. However I have several concerns about the methodological rigor of this study and, consequently the interpretation of the results. Major concerns 1. The two …
Comments to Author
Here the authors report the results of implementation of a relatively new microbiome analysis tool, STRONG, to resolve strain-level features in two urinary microbiome datasets. As the field moves more towards species-level analysis through implementation of shotgun metagenomic sequencing and improved amplicon sequencing methods, the resolution of bacterial strains within the urinary microbiome has become of greater interest. The implementation of methodologies capable of identifying multiple strains of the sample species within shotgun metagenomic sequencing datasets on less understood microbiomes like the urinary microbiome is important. However I have several concerns about the methodological rigor of this study and, consequently the interpretation of the results. Major concerns 1. The two datasets (Chicago and Amsterdam) analyzed in this report do not lend themselves towards exploring bacterial strain diversity in the asymptomatic microbiome. There are two main reasons for this: i) The majority samples analyzed are not from asymptomatic patients. The samples from the Amsterdam study are from patients who were culture negative, not asymptomatic. The point of this study was to analyze the utility of metagenomic sequencing for the diagnosis of UTI - nowhere in the report does it say that the patients from the culture negative group were asymptomatic. These patients could have just been under the detection limit for the clinical urine culture method used. Additionally, only two patients from the Chicago study had not LUTS according to the MS thesis which was the only source I could find for this study. The remaining 16 had overactive bladder which has overlapping symptoms with UTI. ii) Identification of strains using STRONG requires constructing MAGs which in turn requires considerable sequencing depth. Both studies used are of insufficient sequencing depth to produce MAGs for the majority of community members and as it seems from the application here in the majority of samples. The purpose of the Amsterdam study was to assess the clinical utility of urine metagenomics and not to use metagenomics to explore urinary microbiome composition. Because of this the target read depth was purposefully low (~2M per sample). In contrast the target read depth for real dataset that was used in the original manuscript reporting STRONG was on the order of 70 million reads per sample. Similarly, it seems as if the average read depth for the Chicago study was around 7M reads per sample if my calculations are correct and it is unclear how many of these reads were microbial versus host reads. I think it is very likely that the read depth of both studies was insufficient to construct the MAGs needed for strain-level identification by STRONG. I understand that the majority of studies in the field use amplicon sequencing, but there are existing published metagenomic studies of the urinary microbiome that targeted greater read depths and included asymptomatic patients that may be better candidates for analyzing bacterial strain diversity within the asymptomatic urinary microbiome. If the authors do not want to consider including an existing published study that may better suited for this analysis, the limitations of the chosen datasets need to be thoroughly discussed and the title, results and conclusions need to be rephrased to communicate more accurately the clinical status of the patients and that the majority were not asymptomatic. 2. The Chicago study has not been published in a peer reviewed journal (the referenced form is a thesis). Therefore, detailed methodological information needs to be included in this manuscript in order for the rigor of this study to be appropriately validated by peer review. For example, it is unclear from the thesis referenced for this study if any controls for environmental contaminants were included in the sequencing runs. Also, key experimental detail about DNA extraction methods, per sample read depth, number of microbial versus human reads, exclusion criteria, clinical history and demographic information, results of taxonomic analysis need to be included in this manuscript. 3. The Amsterdam study should also be more thoroughly described and should include details about average read depth, number of microbial reads per sample/host contamination, exclusion/inclusion criteria, taxonomic composition, etc. The clinical status of the patients included especially needs to be clarified here, which is that these patients were culture-negative and not asymptomatic. 4. Please clarify what the error bars signify in each figure (1-3). There is no indication in the figure legends. Were sequencing replicates performed or was the in silico analysis performed multiple times? 5. In order to rigorously assess the ability of STRONG to assess strain level diversity in these samples, the authors need to provide taxonomic profiles for the samples. If unavailable, these can be easily generated from the publicly available read data using publicly available tools (Metaphlan, Kraken, etc.). This comparison is essential to be able to assess how well STRONG performs in these samples, especially since bacterial species were only identified by STRONG in 29 of the samples and only 5 samples contained more than one species. This is likely because the read depth of the samples was insufficient to allow STRONG to robustly identify species and strains. The apparent low diversity of the samples could also be due to the fact that the majority of these patients were in fact not asymptomatic. Either way species level profiles generated by STRONG need to be directly compared to those identified by methods more suited to species level identification shallowly sequenced samples. 5. The survey of the literature relevant to this study and its findings needs to be expanded. There are multiple relevant papers that are not cited or discussed, and this has likely led the authors to incorrectly state that "species diversity in the urobiome has yet to be explored" in the abstract. There are multiple published studies that have conducted species level analysis of the urinary microbiome. 6. The methods described in the thesis cited for the Chicago study do not indicate that a water sample or some other environmental contaminant control was sequenced. The Amsterdam study states that a negative sequencing control was included but it is unclear if the data were decontaminated or if the authors of the current study analyzed the negative control to identify contaminants. Because the urinary microbiome is low biomass, it is essential to always include a sequencing control or controls to help identify and remove environmental contaminants. Because neither the Chicago or Amsterdam datasets seem to be decontaminated, the author's results include many likely environmental contaminants. Both Ralstonia (Figure 2C) and Psychrobacter (Figure 3E) have been identified as environmental "kit" contaminants (PMC4228153). This brings in to question their true relevance to the urinary niche and I think they would either need to be cultivated or shown to be absent in sampling controls to begin to suggest they may be part of the urinary microbiome. Moderate concerns 1. Why were only multiple strains of E. coli included in the Mock communities? If the goal of this study is to assess strain level diversity in the asymptomatic urinary microbiome, wouldn't it be more representative to include multiple different strains of Lactobacillus crispatus or other species that are much more prevalent in the microbiomes of women without LUTS? All of the strains included in the mock communities except S. epidermidis are associated with urinary tract infection, so these mock communities would be more informative for assessing the ability of STRONG to identify strain level diversity in patients with UTI or with ASB perhaps than "healthy" women. 2. What read depth was targeted for the mock communities? This is important information that needs to be included for the reader to be able to assess if they inability to resolve strains was due to a deficiency in STRONG, close genomic relatedness, or a limited read depth for high quality MAG assembly. 3. For the Chicago study, were the human reads removed from these samples? Please clarify in methods. Human reads should be removed from read data publicly deposited unless express consent was given by the subjects to have these data public. Also, because this dataset has not yet been published in a peer-reviewed article, a declaration of IRB approval and informed consent need to be included in the methods. 4. The authors state that the sampling method from the Amsterdam study is unknown. The methods section of the Amsterdam study states that they were collected by standard operating procedures from patients attending inpatient or outpatient departments. It is likely that a mixture of catheterized and clean catch samples were collected, but unless the patient already had a catheter it is likely that most samples were clean catch because I believe this is usually SOP for collecting urine to be sent for clinical urine culture for suspected UTI. It might be worthwhile to reach out to the authors of the Amsterdam study to clarify if the sampling method is important for the conclusions of the study. 5. Line 163-164 - I do not believe there is sufficient evidence in the literature to state that E. coli is a common constituent of the female urinary tract or that it is associated with urinary tract health. Just because it can be found in some women without clear LUTS does not mean it's a common member of the healthy female urinary microbiome. ASB is not well understood and there are not sufficient longitudinal studies to conclude that these women with E. coli in their urinary microbiomes never go on to develop UTI symptoms. Minor concerns 1. Line 94: I believe the version number of the SRA toolkit needs to be filled in still. 2. Please include project numbers for both studies to make it easier for the readers to find the data and metadata. For the Amsterdam study, the authors should also include the MS numbers (eg MS_0007) referenced in the original manuscript to allow the reader to better compare results - perhaps in table S1.
Please rate the manuscript for methodological rigour
Poor
Please rate the quality of the presentation and structure of the manuscript
Poor
To what extent are the conclusions supported by the data?
Partially support
Do you have any concerns of possible image manipulation, plagiarism or any other unethical practices?
No
Is there a potential financial or other conflict of interest between yourself and the author(s)?
No
If this manuscript involves human and/or animal work, have the subjects been treated in an ethical manner and the authors complied with the appropriate guidelines?
No: The "Chicago" study has not been published so the authors must include a statement of informed consent and IRB approval.
-
