Filling the Gaps in the Cyanobacterial Tree of Life—Metagenome Analysis of Stigonema ocellatum DSM 106950, Chlorogloea purpurea SAG 13.99 and Gomphosphaeria aponina DSM 107014

Round 1

Reviewer 1 Report

The Authors want to fill "gaps in the cyanobacterial tree of life" and highlight the hidden microbic diversity. The paper is fluent and based on a huge work, however some amelioration can be done. I suggest the follow:

Introduction:

Page 2, line 79: "All cyanobacterial MAGs showed a completeness of more than 90% and a contamination level of below 2%", I suggest to write: "All cyanobacterial MAGs showed more than 90% of completeness and a contamination level below 2%";
Page 2, line 85-86: please rewrite the sentence, it is not very clear to me;
Page 2, line 87-94: please clarify. It not very clear to me the goal of the paper. Want identify the associated bacteria or the cyanobacteria species alone?
I would revise both introduction and abstract to better highlight the scope of the paper. There confusion about the choice and need for axenic or not axenic strain that I guess is you subject of study.

Materials and Methods:

4, line 118-119: "The 118 final metagenomic libraries ranged in sizes from 650 bp to 700 bp.", this is a result;
paragraph 2.3.2: I suggest to insert a flux diagram of the pipeline;
paragraph 2.5: where didi you took the 213 cyanobacterial genomes to do phylogenetic analysis?
“Text mining with all contigs via BLASTN” describe better the procedure.

Results and discussion:

I suggest separating the results from the discussion. The Author instructions of the journal state separate results and discussion paragraph with the results being “a concise and precise description of the experimental results, their interpretation as well as the experimental conclusions that can be drawn”.

Merge subparagraph 3.1.1 with 3.1.2;
In 3.1.1 I would eliminate the numbering of the figure 1 description, alias “First”, “Second”, etc.; similarly all along the results section;
line 172-177, 179-183, 185-195, 209-216 are more appropriate in the discussion as they are not supported by previous publications and not by a figure experimentally produced for this study;
line 222-223: “The generation of Metagenome-Assembled Genomes (MAGs) essentially based on our 222 established binning approach” put it in the methods;
transform the paragraph 3.3.1, 3.3.2, 3.3.3 in a table to better follow the results description;
paragraph 3.7: put it under discussion section.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

This manuscript by Marter et al. describes the genome sequencing of three cyanobacteria, selected to fill gaps in the phylogenetic tree. There is also interest in the composition of the “phycosphere,” which represents heterotrophic bacteria that exist in some sort of consortium with the cyanobacterial phototroph/primary producer. Unfortunately, the manuscript is in the form of a preliminary lab report, containing much data that is just not solid enough for publication. The primary problem is that the assembled Illumina sequencing data has multiple binning problems and the authors’ attempts to correct these are not convincing. A logical approach would be to try different assembly and binning programs to help find artifacts and then make the decision on which data are useful and which are just intriguingly interesting, but not solid enough to publish.  

But, at root, the methods being used have been superseded. Something like the long-read sequencing used in Driscoll et al., Standards in Genomic Sciences (2017) 12:9, is closer to the state of the art, and that study reported three completed genomes in association with a non-axenic cyanobacterial culture.  It must also be said that a quite bizarre DNA extraction procedure was used; I would not trust data from such a technique in my own lab, although clearly sequences were obtained. But there could easily have been differential loss or damage of DNA from some bacteria present and not others, since the cells were exposed overnight to 37C, but presumably cultured at 15-20C and not adapted to 37C.

The manuscript has very long blocks of text that convey little information. And in the end, there is no analysis of the genome content of the three cyanobacteria, although this seemed to be the motivation for the study. The information will require thinning and rearrangement to make a readable manuscript. For instance, relative abundances of bacteria surely provide some indication of their importance in the system, but this information is only revealed quite deep in the paper, well after a reader feels the need for that information in making sense of the binning problems.

A further significant problem is the lack of a description of the history and preparation of the cultures. This is surely critical in assessing whether the heterotrophs are functionally connected in some community with the cyanobacteria, or just accidental free-loaders.

Some specific comments follow, but only for the earlier part of the MS:

Line 64: the binning approach does not need the type of introduction given here, since it is established and not novel; if anything, it will become less important as we rely more on long-read sequencing; CheckM (Ref 22), referenced here, is not a binning program

Line 66: “astonishing accuracy” over-emphasizes the technique; it is of course valuable, but has pitfalls as seen in the inability often to assemble fully complete genome bins

Line 69: Ref 24 is not appropriate here. Ref 22 is and there could be others.

Lines 75-79; should include Driscoll et al. ref

Lines 82-84: I don’t believe the literature agrees that such specificity is necessarily widespread at all

Lines 98-101: explain the meaning of SAG here, as it can mean single cell-amplified genome

Table 1: please use completeness/contamination estimates instead of the subjective “high quality” used to describe the MAGs

Line 113: the use of an extraction kit designed for plant or soil samples is usual and seems more logical. Why was a mammalian tissue extraction kit used? Is this a good recommendation for others to follow?

Line 114: this is a very odd and likely problematic protocol; it is generally recommended to expose extracts as quickly as possible to the stabilizing components in extraction kits. In this case, many cells would be lysed overnight and components exposed to damage.

Lines 118-119: what is 650-700 bp? Insert size?

Line 153: the section on manual curation appears to involve some arbitrary decisions: what were the coverages for the removed contigs in relation to cyanobacterial and other bins?

Line 177: please add a higher magnification of the characteristic heart-shaped cells that are mentioned, as that cannot be distinguished

Lines 199-200, Fig. 1: please add the size info for scale bars directly onto the photographs

Lines 203-204: please annotate the photographs directly, identifying the indicated cell types and associations with matrix, sheath, etc. mentioned on the following lines; it will require larger photos to convey this information properly

Lines 203-204 and elsewhere: indicate how the cyanobacterial cells were collected for this SEM analysis and for DNA extraction; was there any filtration, washing…?

Lines 213-214: how do you know that these bacteria are derived from the original sampling and do not include contaminants introduced during those decades of culturing?

Lines 231-234: the meaning of this sentence is unclear

Line 272: explain what is meant by “deviant nucleotide composition”

Line 261, section 3.3.1: refer to comment above for line 66, since all sorts of problems are emerging here. Please discuss the parameters being used for binning. Why haven’t other binning programs been tried for a complementary approach, e.g., MetaBAT?

Lines 288-293: it is not valid to propose a new classification for an organism based on one or two genes in a 1.5 kb contig, because such DNA fragments can easily be shared horizontally and the home genome cannot be known.

Lines 261-302: the material discussed here is in a preliminary form that is not ready for publication. Decisions on which portions of data might be publishable should be made.

Lines 346-397, section 3.3.3 on text mining: this reviewer is not convinced of the usefulness of this approach other than helping to guide making initial sense of data

Line 398, section 3.4.1: the first priority in discussing the metagenomic bins should be to order them in some way, primarily by sequencing coverage as a proxy for relative abundances. The lower the coverage, presumably the lower the relevance, as there are likely contaminations in the system. The sample preparation and history of the cultures should be better discussed, as these are important in considering whether the bins are functionally relevant to the cyanobacteria, meriting designation as phycosphere members.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf