the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 Aug 2026
Supraglacial microbial communities from an isolated high mountain glacier in Türkiye
Abstract. The biodiversity of glaciers has been studied in polar regions and mountain ranges, yet isolated high mountain peaks covered by glaciers in temperate regions – islands of ice – are virtually unknown. These systems are particularly at risk of rapid change and disappearance as temperatures increase, thus understanding their biological diversity and the potential for organisms to survive as the landscape alters is time-critical. In this study, we assessed the community composition of a glacier surface in Türkiye, located on the extinct volcano Ağri Daği (Ararat Mountain). We collected samples of sediment (cryoconite) from water-filled depressions (cryoconite holes), then analysed samples through culturing, microscopy, metabarcoding (16S and 18S rDNA), metagenomics and metatranscriptomics. We cultured and identified 158 bacterial strains belonging to 11 genera (Cryobacterium, Polaromonas, Flavobacterium, Arthrobacter, Massilia, Devosia, Paeniroseomonas, Deinococcus, Glaciihabitans, Knoellia, and Rugamonas) and 41 species. Nine species of Cryobacterium, Flavobacterium and Knoellia did not grow above 10 °C, indicating particular vulnerability to climate warming. Amplicon-based community structure was dominated by Cyanobacteria, with the genus Tychonema particularly abundant, while cultured isolates represented only a small fraction of amplicon diversity but captured phenotypically diverse, cold adapted heterotrophs. Functional potential in metagenomes revealed that cyanobacteria dominated bacterial primary production, whereas other phyla specialized in nutrient scavenging, which was further apparent in metatranscriptomic analysis. Transcriptomic analysis demonstrated disproportionate dominance of Cyanobacterial transcripts across core metabolic pathways. Amplicon analysis of 18S rDNA and rRNA reads from metatranscriptomics showed that eukaryotes were active, with communities dominated by Ciliophora, while metazoan grazers identified in other cryoconite systems (such as tardigrades and rotifers) were not detected in either molecular data or microscopy, suggesting a simplified, protist-dominated food web. The combination of different microbiological and molecular approaches demonstrates that this rapidly changing glacier hosts cyanobacteria-dominated, bacterial–protist assemblages with absence of metazoan grazers, showing that isolated low latitude glaciers are a viable microbial habitat under threat.
Competing interests: At least one of the (co-)authors serves as associate editor for the special issue to which this paper belongs.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(1646 KB) - Metadata XML
-
Supplement
(9835 KB) - BibTeX
- EndNote
Status: open (until 22 Oct 2026)
- RC1: 'Comment on egusphere-2026-4517', Anonymous Referee #1, 30 Sep 2026 reply
-
RC2: 'Comment on egusphere-2026-4517', Anonymous Referee #2, 02 Oct 2026
reply
Comments to
“Supraglacial microbial communities from an isolated high mountain glacier in Türkiye”
General comment
The manuscript “Supraglacial microbial communities from an isolated high mountain glacier in Türkiye” offers meaningful insights on glacial microbial ecology on multiple fronts, targeting both prokaryotes and eukaryotes communities thriving in cryoconite holes. Particularly relevant is the studied glacier being a “temperate” one, as this class is generally underrepresented in the existing literature. This knowledge gap is methodologically addressed by employing a multi-omics approach that includes cultivation of isolates, amplicon sequencing, metagenomics and metatranscriptomics. This clearly adds to the solidity of the conclusions drawn, although some clarifications and improvements are needed, mostly for the bioinformatic/biostatistic treatment and the documenting of sequencing results.
Minor formal amendments are otherwise suggested, with the composition barplots showing the biggest room for improvement.[#]. [Line referenced] Comments:
1. [16] The glacier surface has been extensively described as a multi-matrix reality, where microorganisms are able to colonize a plethora of different habitats (snow, ice, surface sediment, cryoconite sediment…). The phrasing “… we assessed the community composition of a glacier surface…” implies the representativeness of cryoconite communities for the broader ecosystem, which is not quite realistic.
Suggestion: slightly rephrase to provide a better framework for the manuscript.2. [18-19] The same metabarcoding techniques have been referred to in different ways across multiple instances in the manuscript (“16S and 18S rDNA” in lines 18-19; “18S rDNA” in line 27; “16S rRNA gene sequencing” in line 118; “16S rDNA amplicon sequencing” in line 245; “16S rRNA gene sequencing” again in line 359).
Suggestion: keep wording consistent (e.g. 16/18S rRNA gene amplicon sequencing).3. [51] They also constitute micro-environments relatively protected from direct UV radiation, a major player in DNA/RNA degradation.
4. [65] Other glaciers from the same mountain range or different ranges? In this context, this notion is probably worth discussing briefly.
5. [81-82] Here, the study area is defined "an isolated mountain peak in continental climate". In L86 "an isolated temperate mountain peak", and a "subtropical area" later in L335, resulting a bit confusing.
Suggestion: clarify the proper characterization and refer to it in a consistent fashion. Meteo/climatic data might help if any claims are made.6. [87] The (ii) aim is mentioned here, but it does not seem tied to any of the conclusions.
7. [101] While cryoconite holes are generally stable environments and comparable to each other within the same glacier in terms of community composition, having only 3 as representatives for the whole glacier might raise some concerns. What were the reasons/limitations behind this sampling strategy?
8. [101-102] "Chipping frozen cryoconite” implies daily freeze-thaw cycles if not the absence of liquid water altogether, which is a distinctive trait of cryoconite holes. This might lead to the conclusion that the results inferred from RNA expression might underestimate the true activity of these communities or be somehow skewed in favor of those microorganisms particularly adept at keeping the freeze-related stress under control.
Suggestion: clarify the conditions of the cryoconite holes sampled in the manuscript.9. [132] The reference publication for BioEdit is missing.
10. [133] The information on the BLAST Nucleotide Database variant used to retrieve the taxonomies is missing, along with the database version identifier and the BLASTn suite citation.
Suggestion: include database identifier (“nucleotide-nt” / “core_nt” / …), database version (1.1 or at least mm/yyyy of the BLAST search) and a citation to the BLASTn suite (Zhang, Z., Schwartz, S., Wagner, L., & Miller, W.: A greedy algorithm for aligning DNA sequences. Journal of Computational biology, 7(1-2), 203-214, 2000.)11. [134] It’s great that the isolates’ sequences are already publicly available. No availability statement for the Illumina sequences (amplicons/MAGs/transcripts) can be found in the manuscript though.
Suggestion: add data availability statement for Illumina sequences (NCBI accession/upon request…).12. [137] The original amount of cryoconite used for the extractions (but also organic content, microfauna and granule size analyses) is missing.
Suggestion: include the amount of sediment extracted/processed (grams/wet weight/…).13. [140-141] Clarify that the extracted DNA concentration in the samples is the subject of the pre-PCR normalization (not the sample itself), along with the hypervariable regions targeted by the primers used.
Suggestion: rephrase sightly and add hypervariable regions information in the manuscript (16S V3-V4, 18S V4).14. [151] Please clarify whether the pooling was conducted “sample-wise” on the reaction triplicates mentioned in line 142 or among samples (consequently losing single cryoconite information). Was each cryoconite sequenced separately?
15. [154-156] The referenced paper by Ezzat and coauthors (“Molecular modeling analyses of functionalized cellulose”) does not include sequencing or a sequencing protocol. Versions and references for Deblur and DADA2 are missing.
Suggestion: double-check the reference and/or clarify (in the paper or as part of the supporting information) whether the bioinformatics were run with default tool parameters or if any flag was manually set (in particular, the read quality filtering threshold and the read length, an important parameter for Deblur). Add relevant versions and citations.16. [169] The link to the pipeline repo (https://github.com/michoug/MAGsGeneration) leads to a GitHub 404.
Suggestion: check the link functionality.17. [170] For reproducibility purposes, please clarify whether each cryoconite sample was assembled independently or pooled/co-assembled.
18. [178] Release version of GTDB-Tk and relative database version from which the taxonomies were retrieved and citations are missing.
Suggestion: include GTDB-Tk release, GTDB (database) version and citation.19. [196] Generally, adapter and quality trimming are conducted prior to sorting RNA into ribosomal and non-ribosomal reads. Please clarify the reason behind this strategy.
20. [202] Citation to R Statistical Software is missing.
21. [206] Is the reported uncertainty a std error or a std deviation? Please clarify.
22. [Figure 3] Might benefit from being divided into two vertical columns or potentially a landscape orientation.
23. [239] The section presents the relative abundance of the top five most abundant phyla, but it lacks sequencing and data treatment documentation, which would improve its backbone and the reproducibility of the results. This also partly relates to comments #14 and #15.
Suggestion: include (in the supporting information) overall/sample-wise sequencing yield from which the relative abundances were computed, as well as whether any data pre-processing was conducted (e.g. presence/absence/prevalence/singleton filtering or rarefaction etc.).24. [242] The percentages presented add up to 92.9%. Is the remaining fraction unclassified at Phylum rank or an unspecified group of “residual” Phyla?
25. [245] Figure 4 could see some improvements in terms of content and readability. Here, the vertical axis label (“Mean Relative Abundance”) hints at an average composition, although single-sample information is never provided (again related to comment #14). The plot subtitle is redundant with the caption, and the palette used can be very problematic for a color-blind reader, especially given the 20-item legend. The “Verrucomicrobiota” phylum label also seems to be cut due to the plot size being too small. Explicitly setting the y-axis interval from zero to 30% would also help the reader make better sense of the proportions displayed, instead of an “open-end” scaling. Having the phylum name below the corresponding stacked bar would also simplify the overall figure.
Suggestion: if possible, provide a stacked barplot with the three samples on the horizontal axis, either including it directly as a compound figure, or at least as a supporting figure. Fix the color palette, avoiding having the same color coding for different genera. Consider plotting less items at higher rank if the information loss is deemed acceptable (e.g. at family rank).26. [249-250] This statement fits better a method section, while the whole section needs more details about how the comparison was conducted. Was a subset of the isolates re-extracted and Illumina-sequenced or were the sanger sequences simply aligned to the ASVs produced as detailed in section 2.4? How were the methodological discrepancies addressed (if any)? Also, double-check singular-plural consistency in “16S rRNA gene […] were sequenced”.
27. [259] If possible, include in the supporting information the details about the retrieved MAGs, in accordance with the “Minimum Information about a Metagenome-Assembled Genome” guidelines, along with the corresponding GTDB taxonomies (the report of Checkm2 is a great place to start). This is relevant information to frame the results and discussion and becomes crucial when comparing the work to other investigations conducted in similar environments.
Suggestion: include a supporting table containing sample ID, MAG id, completeness, contamination, coding density, contig N50, genome size, coverage, total contigs and full GTDB taxonomy (if retrieved) at least for the MAGs with completeness > 50% && contamination <10%, highlighting the subset considered for the downstream analyses.28. [264] The generation of the NMDS presented needs to be better documented.
Suggestion: include (either in the methods/caption or supplementary information) the nature of the underlying data (global/per-MAG occurrence/counts?) and how it was handled (any data transformations applied? What is the underlying distance/dissimilarity statistic of choice?), along with how the PERMANOVA was performed.29. [Figure 7] The same suggestions as Figure 4 on content and readability apply. Please refer to comment #25. In addition, the legend is only comprised of 16 items instead of 20.
30. [313] Methods section 2.6 mentions pre(bacterial)- and post-sequencing rRNA depletion, but the retrieval of taxonomies from ribosomal reads is never documented.
Suggestion: include in the methods the relevant workflow.31. [325-328] While cryoconite hole communities are certainly not comparable to the microbiome commonly found in (e.g.) soils, describing them as a “relatively simple” on the sole basis of the fact that “they are dominated by four phyla” might be a bit of a stretch. In its current version, Figure 4 might contribute to this misconception, since a high number of ASVs is grouped into a single placeholder category (“other”) despite consistently representing a large fraction of the corresponding phylum.
Suggestion: re-phrase the statement or substantiate it with further arguments. Computing alpha-diversity metrics (observed, Shannon, (inverse) Simpson) and comparing them across the study areas discussed could be one way to go. Rarefaction curves might also be of help in determining the efficiency of amplicon sequencing in terms of taxa discovery. In general, this section addresses only prokaryotic diversity: renaming it as such would improve the overall structure.32. [387] This substrate similarity to a glacier in Alaska is quite striking. What could be the driving factor/s?
33. [429-430] These conclusions are derived from the analysis of three cryoconite holes, and it cannot be excluded that part of the true diversity might have been missed. This fact is never mentioned in the manuscript.
Suggestion: include a “cautionary” statement about the limitations of the sampling design.34. [451] Is there a reference no. for the exploration visa to be added here?
General form: the manuscript in its current version needs some careful language edits. Some sentences do not flow naturally (e.g. L39 “increased”, L361-362 “with […] were” statement) or need to be checked for grammar consistency (e.g. L210 leftover “a” in line with the figure (left), L410 “From other hand”, Supporting Table 1 header “Accesion”).
Citation: https://doi.org/10.5194/egusphere-2026-4517-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 192 | 114 | 32 | 338 | 40 | 26 | 35 |
- HTML: 192
- PDF: 114
- XML: 32
- Total: 338
- Supplement: 40
- BibTeX: 26
- EndNote: 35
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
I generally appreciated the paper by Kowalik et al., and I think they provide interesting data that deserve to be published. Anyway, some adjustments are required.
Generally, I would suggest the authors to be more careful when they discuss and combine sequencing data with isolation data and I think that some methodological aspects need to be clarified.
L25: cyanobacteria normally is written with the first capital letter in the manuscript, I would make it uniform (it occurs also in other part of the text). Furthermore, “were specialized” maybe?
L39: “increased AND demonstrated”?
L55: “Combination OF”?
L55-56: “metatranscriptomics and culturing of organisms ACROSS a range of temperatures”?
L68: I think it should be “Oscillatoria taxa”.
L101-102: Were cryoconite fully frozen then? No liquid water was available? Could you specify it? In case I would modify the abstract “water-filled depressions”.
L101-104: Did you use anything to better preserve samples especially for RNA analyses?
Fig.1A: I think it would be more interesting to have a map of the Ağri Daği volcanic complex, instead that its location on such a large-scale map. Maybe you can give more importance to panel B.
L119: I think the format of the citation should be fixed (same for other citations below).
L130: There are two closing brackets after Bio-Rad.
L140-141: I would specify the hypervariable regions you are amplifying.
L154: I would specify the filtering parameters for both amplicons.
L170: “were removed”
L185: Did you check for DNA contamination and eventually treat the samples with DNase?
L199: I would specify that the MAG catalogue was created with the MAGs reconstructed from metagenomic analyses.
L202: I would appreciate more details about the statistical analyses.
L210: There is an “a” next to the image, I think by mistake. To me it is not fully clear what area the arrows are pointing. Maybe a circle that underlies the extension of the granule would be clearer.
L214: I would say “varied in colour” instead of “were pigmented”, just because then there is “colourless”.
L218-224: I would be more careful, since some identity percentages are very low. I would not say that these bacteria are assigned to a species, but I would say that the “best hit” or the “closest match” were..... And I would also mention that some generally did not have good matches in databases. Maybe you can also just mention the genus, but not the species. For the same reason, I would remove Fig. S2.
L250-251: I would add the total number of isolates you obtained “In total, the N isolates corresponded to only 21 of the 511 ASVs”
L253: “accounted for only 1.86% of the relative abundance”?
L254: I don’t know if I would mention it. For the same reason I already mentioned above. The identity threshold is set slightly too low. I would also adjust the discussion about this aspect (es. L342, L359)
L255: I would add the number of these ASVs.
L265: I think that you can keep only Fig. S4, it is clear enough and provides all the data, with no need to have both Fig. S4 and Fig. 5.
L277-278: This sentence was already present in L272-273.
L284: Can you confirm that the genes you found with metatranscriptomic analysis can all be found in the MAGs you reconstructed from metagenomic data? None of them come from other potential genomes that were not reconstructed? If yes, I would rephrase the sentence to make it clearer, so that the next one will also be clearer.
L325: I would specify bacterial phyla just because you analyzed both 16S and 18S rRNA.
L333: “Among which”?
L345: “recover only a fraction”?
L346: I think you should discuss the OTU threshold limit better. I am still convinced that 97% may not be the best threshold in such a context.
L355: “Successful” instead of “representative”?
L356: I think this sentence is too speculative in the present version. Especially because with your isolation strategy you could not detect isolates growing at temperatures >10°C exclusively.
L410: “On the other hand”?