A Comparison of Structure Determination of Small Organic Molecules by 3D Electron Diffraction at Cryogenic and Room Temperature

Round 1

Reviewer 1 Report

In the manuscript entitled “A Comparison of Structure Determination of Small Organic Molecules by 3D Electron Diffraction at Cryogenic and Room Temperature” Taimin Yang et al. describe the differences of structure solutions of organic samples from 3D ED data collected at different temperature conditions.

This review is quite short but the topic will be of interest for the readers of this journal. Unfortunately, authors were not able to present their research in a reasonable way and proof all statements claimed in the abstract. The Introduction is poor and very confused and in Results several absurd and unexplained values are reported, that question the entire structure solution from such experimental data. The paper cannot be published in the present form and must be widely re-checked.

 

General Comments:

The Results chapter is frankly not understandable for anyone that is not deeply inside the topic. Please, expand and specify better important values that appear in tables. It is not clear whether authors used merged data for all examples.

Both cif files correspondent to the ABTC structure do not contain H-atoms. This is inconsistent with claims in the text and with figures.

Moreover, the necessity of merging (in different ways for different data sets) complicates the comparison of structure solutions obtained at different temperature conditions, which is supposed to be the main focus of the paper.

Finally, 3D ED, MicroED, cRED and other acronyms are used through the text, apparently indicating the same methodology. Authors should check carefully this point in order to give precise information to readers.   

Detailed Comments:

• Line 13-15

3D electron diffraction (3D ED), also known as micro-crystal electron diffraction (MicroED), is a rapid, accurate and robust method for structure determination of small organic molecules from micron- and nano-sized crystals.

3D ED method is not only for structure determination of small organic molecules, but of different kind of materials and small organic molecules is only one of them.

 

• Line 26

I/?

 

• Line 41

Dorset et al. (according to signed Ref.4)

 

• Line 47

…Kolb’s group and Zou’s group…

Generally, this is not a proper way to cite previous papers. Better to use “et al.” or “with co-workers”.

 

• Line 50-51

The development of these two methods allowed researcher to determine a large number of inorganic structures from nano- and submicron-sized crystals…

Please, add references!

 

• Line 51-52

…more than 100 structure in our lab alone.

This part of sentence should be removed or rephrased. Authors may present a statistic among the electron crystallography groups using 3D ED methods, but not only one group alone. “Our” should not be used at all, since this manuscript is written by two scientific groups.

 

• Line 55

…using a hybrid detector in 2016.

Is it a first time a hybrid detector was used for the structure determination? Please, specified it. Also, it would make sense to add some words about advantages of using such detector and how it improves structure determination of organic samples especially.

 

• Line 55-57

Recently, several articles [12–15] describing the routine and rapid structure determination of small organic molecules using 3D single crystal electron diffraction were published.

This sentence is not well connected to the previous and the next one. Why was selected this group of references? Any particular way of data collection? Additionally, more recent papers should be added.

 

• Line 59

Later he determined…

Who is “he”? There are three authors in the correspondent reference.

 

• Line 62-63

…by several groups; Gonen’s group [18,19], Yonekura’s group [20], and Abrahams’ group [21,22].

This sentence should finish after “several groups” with following citations. Add more references.

 

• Line 63-64

3D electron diffraction methods can also be performed in nanobeam mode [23–25] and combined with 4D STEM…

In which mode were performed previously mentioned 3D ED structure. In which paper 3D ED was combined with 4D STEM? What authors really mean here for 4D STEM?

 

• Line 65-66

In our publication … we showed…

This sentence should be rephrased, avoiding using “our” and “we” referring not to the present manuscript.

 

• Line 66-67

…the structure model can be improved by 3D merging electron diffraction data…

Reference 28 is not unique that support this statement. Add Brázda et al. (2019) Science; Woollam et al. (2020) CrystEngComm; Hamada et al. (2020) Bull. Chem. Soc. Jpn. …

 

• Line 68-69

Furthermore, our recent results showed that it is possible to solve unknown protein structure by MicroED…

“Our” should be replaced with “some”. Add more references for the unknown protein structure solved from 3D ED, for example Lanza et al. (2019) IUCrJ.

 

• Line 69-71

…and visualize ligand binding interactions in human carbonic anhydrase isoform II (HCA II) with acetazolamide (AZM) [31].

This sentence should finish with “and visualize ligand binding interactions”, further detailed information is not of interest of current paper.

 

• Line 71-78

We have also demonstrated that accurate structure models (average deviation from the reference model: 0.03 - 0.07 Å) can be obtained rapidly and routinely from electron diffraction data of inorganic and metal-organic crystals based on kinematical approximation, despite the high R-values (15-30%) due to dynamical scattering [32,33]. Our group developed software Instamatic to facilitate the 3DED/MicroED and serial electron diffraction collection scheme [34,35]. Based on Instamatic, we further developed several high throughput and automatic data collection and processing protocols and collect them into software packages [36].

This part of the Introduction looks like a self-advertisement and highlights details that are not directly connected to the main topic of the paper. I would strongly recommend to remove it or at least rewrite with respect to other scientific groups, that are working in the field of electron diffraction. 

 

• Line 82

…two organic crystals…

Which exactly crystals? Why authors select these examples? Add shortly some details about examples and how they were selected?

 

• Line 83

…continuous rotation electron diffraction (cRED) data…

In the above part authors introduce different methods and modes of 3D ED data collection and only know introduce method that was used in current paper. Some additional info about cRED should be added.

 

• Line 85

…and data merging…

Why authors deviate from the temperature effect to the merging data? It is completely different things that should be discussed separate.

 

• In Materials and Methods is not mentioned where samples were purchased or how they were crystalized.

 

• Line 93

Abbreviation SAED was not introduced.

 

• Line 88-89

…(H₄ABTC, 358.3 Da)…

In the abstract authors refer to the second examples as ABTC and here as H₄ABTC.

 

• Line 113

SHELXL

 

• Line 113

…using only five restrains.

Exactly, which restrains were used?

 

• Line 129-130

Add explanations why the R_int value may increase significantly by data merging.

 

• Line 134

Table 1

 

• Line 159-160

…in which 20% were from the Zr-MOF, 75% were from the H₄ABTC crystals…

And what about last 5%?

 

• Line 162-163

Also the Zr-MOFs was tested as catalysts for the oxidative coupling of arenes via C–H/C–H activation[46].

This sentence and correspondent reference should be removed since Zn-MOF is not of the interest of this paper.

 

• Line 187

I/σ

 

• Line 199

I/σ

 

• 1E

H-atoms and some intermolecular distances are missing compare to the Fig. 1C and 1D.

 

• 3C

Typical electron diffraction pattern of Zr-MOF crystal collected at room temperature. We note that the MOF crystal had larger unit cell and diffracted to higher resolution.

This information should be removed since Zn-MOF is not of the interest of this paper.

 

• Capture to Fig.4

…difference Fourier maps…

 

• Table 1

Chemical formula and weight are missing

Units of cell parameters are missing

What means values with *?

No. of parameters should depend only on the structure and not on the way data were collected. Why there is a difference in these values?

No. of restraints is also different, while for an accurate comparison between data qualities it would more correct to use the same restraints.

R_int of single data set is 6%, while for merged data is above 50%. These values are very anomalous. One should conclude that authors are merging unrelated things. Moreover, it is surprising that with a so high R_int they still get structure solution ab-initio.

 

• Table 4

Chemical formula and weight are missing

Units of cell parameters are missing

What means values with *?

No. of parameters should depend only on the structure and not on the way data were collected. Why there is a difference in these values?

No. of restraints is also different, while for an accurate comparison between data qualities it would more correct to use the same restraints. For 100K this value is 0?

 

Why for sucrose a column with an apparent non-merged data appears, and the same is not true also for ABTC?

Comments for author File: Comments.pdf

Author Response

Please see attached pdf file

Author Response File: Author Response.pdf

Reviewer 2 Report

Page 1 line 23 “ABTC” is an undefined acronym

Page 1 line 26 indicators list in incomplete.

Page 2 line 33 Electrons interact with matter 10^6 times stronger than x-rays, this is a meaningless statement without context, for which quantity of x-rays and electrons, the interactions vary significantly with energy. Gives values and cite a source if how this was determined, I have seen values of 10^3 up to 10^6 but without context this is a meaningless statement.

Page 3 line 86 “researchers to setup experimental conditions” is a clumsy phrasing, I would recommend expanding on this and articulate what the benefits to the experimenter would be.

The quality of data from electron diffraction experiments is highly reliant on the thickness of the crystals being collected from, and there is no mention of the crystal thicknesses used in these experiments. If there is meaningful comparisons between the room and cryo temperatures has to be made, then accounting for the crystal thickness must also be taken into account. This manuscript does not attempt to take into account the varying thicknesses of the crystals used to collect data. The data is also merged from several crystals. Again no indication of the thicknesses and morphologies of the crystals is taken into account by the study, which can have a significant outcome to the claims made by this study. Given the result of the manuscript are counter to the current thinking within the field, then substantial evidence needs to be brought to bear to demonstrate the perceived wisdom is wrong. The submitted manuscript does not present data which has been analysed sufficiently to run counter to the current assumptions within the field.

According to table 1 there are five fewer data sets merged for the cryo data set than for the room temperature. For a fair comparison, they should be the same, this may well account for the lower resolution, number of unique reflections. It is also unclear how you can have fewer reflections and have higher data completeness in the dataset.

In terms of the ice rings, is this not a case of collect better data? It’s more a function of not having the proper equipment in terms of no cryo-shield being available or the vacuum in the TEM rather than an experimental data problem? Is it not the case that collecting data shortly after the holder is inserted, fewer or no ice crystals will be present, and that the experiment requires several fresh data sets to be collected rather than all from one grid in the same session? I also fail to see how ice rings and inelastic scattering limit the resolution as these only impact the low-resolution reflections; please explain why this is the case in more detail.

I don’t see the relevance of the Zr-MOF other than to acquire a self-citation; please remove this sentence or elaborate on why it is meaningful in the context of this manuscript.

For the benefit of the reader, why is the sucrose data was only rotated over 60-degrees, when the ABTC was over 110 degrees.

I am not convinced that the evidence provided supports the paper's conclusion regarding the cryo-arguments since the differences in crystal quality and thickness have not been considered in the investigation. However, with respect to the idea more data and better coverage of reciprocal space gives better results, this is hardly a new result, especially given this group has been involved with developing electron serial crystallography.

While I have my reservations about the claims currently, they may well be correct but need more analysis and elaboration within the text.

Author Response

Please see attached pdf file.

Author Response File: Author Response.pdf

Reviewer 3 Report

Introduction

Comment: It is too long, it contains 1/3 of the manuscript text. It can be perfectly shorten by condensing the full description of the authors group achievements.

Line 26: “of” should not be there.

Line 33: I have seen in various publications a factor of 10³ or 10⁴. If this does not apply here, please provide a reference from where this information has been obtained.

Line 36-38: Sentence starting with “It was difficult … higher electron dose”, bring it to present tense.

Line 44: cases with s.

Line 46: or instead of and.

Line 50: researchers, with a final s.

Line 51: nano- and submicron- is the same, decide which one to use.

Line 59: dynamical refinements.

Line 66: take out 3D.

Line 69: structures, with a final s.

Line 75: the Instamatic software.

 

Materials and Methods

Line 91: add an a in front of LaB6.

Line 93: collection should be a verb not a noun.

 

Results and Discussion

Discussion should be without s

Comment: when writing difference map, add Fourier in-between: difference Fourier map.

Line 105: is should be was

Line 105: a should be an

Line 106: after electron dose, of should be added.

Table 1: The table caption should say that the reported values are not only statistics but also parameters. There is an asterisk in the CC1/2 factor, I don’t know what that means. I don’t know what “No. of total refined data” means, I would write “Used reflections” or something similar. Moreover, while this value coincides with the number of unique reflections for sucrose and sucrose (100K), it does not with the sucrose single dataset, why?. The note about brackets do not apply for the unit cell parameters, please specify it. Add the citation that corresponds to the reference model at the end.   

Figure 1: Specify that SHELX was used to obtain the potential maps and what the different potential colours mean.

Line 113: Table 1 says 142 parameters.

Line 114: specify that these are symmetrically-independent reflections.

Line 116: which database? Reference needed.

Line 119: table 2 does not include H deviations. Since it is discussed in the text, it needs to be reported in the table.

Table 2: Add the citation that corresponds to the reference structure model used.

Line 124: add an a in front of single.

Line 125: In figure 1, specify which threshold was used for the display of the different electrostatic potential maps to ensure that the correct comparison is made.

Line 127-133: This number of conclusions is a bit disordered and not really easy to follow. Bring the ones related to the merged data after discussing the obtained averaged bond lengths, and the ones for the single data after listing its found averaged bond lengths. 

Figure 2: The H-bond is between Hs and Os, instead the O-O distance is highlighted. Specify in the figure caption from which data this model is obtained.

Line 139: Since ADPs are discussed, a table of the different ADPs should be provided, at least the isotropically calculated ones if the anisotropic ADPs were refined.

Line 144: If this sentence refers to the anti-contamination device, please add it instead of “noncryo-microscope” wording. The main reason why the data may be worse in cryo-conditions is because of ice + hydrocarbon deposition, which it can be highly minimized by the use of an anti-contamination device, leaving a couple of hours without significant contamination that can be used to acquire 3D electron diffraction data at the same or similar resolution.

Line 151: I would add after “vacuum condition”: and stable under the electron beam.

Line 159: These 52 datasets include RT and cryo-conditions? Specify.

Table 4: The table caption should say that the reported values are not only statistics but also parameters. There is an asterisk in the CC1/2 factor, I don’t know what that means. I don’t know what “No. of total refined data” means, I would write “Used reflections” or something similar. The value at “No. of total refined data” for the 100K structure do not coincide with the “No. of unique reflections” even if the same resolution is taken, why?. The note about brackets do not apply for the unit cell parameters, please specify it.

Table 5: The column “Atom2” in the middle contains values that are not atom labels. Please revise it.

Line 184: Q can be written in lower-case: q. In front of “q-peaks” add the given.

Line 192: Well, I think that some description is missing here regarding the refined structure at 100K as well as the tables for bonds and ADPs.

 

Conclusions

Comment: I would add that cryogenic conditions may be applied when a compound is pretty beam sensitive in order to increase the crystallinity of the material under the electron beam since it is usually used in this way, not because of vacuum conditions.

Line 199: after ratio: if anti-contamination device is not available.

Comments for author File: Comments.pdf

Author Response

Please see attached pdf file.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Ones again, in the Abstract authors refer to the second examples as azobenzenetetracarboxylic acid and in the Materials & Methods as tetra-carboxylic 5,5'-[(E)-1,2-diazenediyl]diisophthalic acid. Can authors select and use the same naming through the whole manuscript?
 
Chemical formula and weight are still missing in both Table 1 & Table 4.
 
In tables about "Bond lengths" (Table 3 & Table 5), third value appears as d(Å) in Table 3 or Length(Å) in Table 5, respectively. Can authors select and use the same naming through the whole manuscript?
 
Conclusions should be expanded putting emphases on the general importance of 3D ED method and possible further advances in the field of electron crystallography. 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

The revised version the manuscript is significantly improved over the original version.

There is still a questions about the conclusions drawn in this study, they seem very counter intuitive and lack supporting evidence. There is also the question of were the data collection, is it collected sequentially on the one grid? Room temperature followed by cryo cooled, or are these completely separate experiments? If sequentially collected the crystals may have experienced radiation damage during the room temperature collection, as radicals diffuse across the grid when damage occurs during the room temperature data collection.  

The hypothesis that cryo gives worse data than ambient temperature goes against decades of work in X-rays, neutrons and electrons diffraction, and imaging studies. Yes ice rings are present, bu they only effect a few low resolution reflections. There is increased inelastic scattering as well, but inelastic scattering is also low angle scattering, and does not effect high angle data. Evidence supporting the effects of the inelastic scattering as a function or resolution shells, comparing cryo and ambient temperatures, would give some substance to the  claims being made. Quantifying the signal to noise of the background around reflections in the different resolution shells and between the reflections would add evidence and credence to the proposed hypothesis, current the manuscript lacks an evidence backed claim about why the resolution between ambient and cryo conditions is observed. 

In the case of protein crystals they do lose resolution with Cryo-conditions, but this is due to the water content of the crystals causing increased mosaicity due to the water to ice phase transition, which would not be the case for these crystals as the do not contain water.

Alternatively you could show that cryo conditions induces some sort of mosaicity in the crystals causing a degradation in the measured reflection intensities would be a suitable argument to prove the point that cryo degrades the resolution of the data. 

 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

Authors in this publication want to prove that merged datasets at room temperature compared to 100 K datasets acquired in non-dedicated cryo-electron microscopes can produce equal or better structure solutions and refinement. Such kind of work is very interesting and I think that is worth for publishing. However, strong flaws are detected which prevent me to suggest such manuscript for publication.

After the first round of corrections, the Introduction and Materials and Methods section is fine but the results and discussion part is a too crowded of values and their small discussion. I would suggest to make two different sections: one in which the authors explain how they solve and refine the structures together with the most important parameters of such analyses, which would be the Results section, and another one where the results are discussed, which would be the Discussion section, since not that much discussion is available now and the aim of the paper requires a large and careful one. Given that the authors come from one of the well-established electron crystallography groups, that should be done and a careful check for consistency across the manuscript must be carried out. At the moment, it seems like every part of the manuscript is edited and written by different persons without consistency. Further comments of the different sections are written hereunder. 

Abstract

Line 15: Submicron- and nano- are the same dimensions. Only using submicron-sized crystals should be enough. I’ve already made this comment in the previous review, albeit in the introduction part.

Line 18: that should be placed in front of 3D ED/MicroED to make the sentence understandable.

Introduction

Line 34: Electrons with an s. Remove is for are.

Line 46: I don’t understand the sentence “electron diffraction data were obtained from incomplete data”. The ED data is already incomplete. I guess that electron diffraction data should be replaced for crystal structure models?

Line 68: I would recommend to remove the citation of https://doi.org/10.1038/s42003-018-0263-8 (Nanoscale mosaicity revealed in peptide microcrystals by scanning electron nanodiffraction) since it does not work with 3D ED data but 4D-STEM maps without tilting the holder in comparison to the other cited paper of Gallagher-Jones et al.

Materials and Methods

Line 105: Most materials are stable under vacuum condition. I understand that zeolites with some organics in the cages may go out of the structure in vacuum and destroy the crystallinity of the particles or in some minerals where the water is trapped inside. However, I don’t see that the stability under the vacuum condition should be the main argument to pick these two samples, a lot of organics could follow such characteristic. Authors can just say that they use this low symmetry crystals for the temperature comparison purposes to avoid the impression that a large amount of materials are not stable under the high vacuum of a TEM. 

Line 113: change selective for selected.

Line 114: add pixel between hybrid and detector.

Line 116: remove directly since some merging and restraints were needed in some cases, thus it could be misleading.

Results and Discussion

Subtitles Sucrose crystals and H4ABTC crystals could be left without crystals. There is no need to add the noun crystals in these subtitles.

                Sucrose crystals

General comments:

• This section is not really well structured and it is difficult to follow. You have to constantly go back to understand if it is talking about the RT or 100 K dataset. There has been an improvement with respect to the first version, but the authors should make it much clearer. Furthermore, the discussion of restraints and parameters while discussing the refinement of the RT dataset is not really explained for the 100 K data or, at least, it is not clear to me. The same parameters, bond lengths, ADRAs, H-bonds, … need to be discussed from both RT and 100K structures for the sake of comparison.
• If the ADPs are negative, it doesn’t make sense to perform the refinement with anisotropic ADPs, keep it isotropic.
• The cifs provided to the journal still do not include the hydrogen positions.

Specific comments:

• Line 120: Remove “We first use sucrose crystals as a testing sample.”. The authors already said that in Materials and Methods section, it’s not needed here.
• Line 124: Taking into account 1.13 degrees/s, 60 degrees of angular range and 0.3 s/frame, results in a total of 177 patterns. I would change the stated 190 for 180.
• Line 137: The authors did not obtain the Hydrogen positions, they just added them according to a chemically senseful model. Please modify this wording to avoid confusion.
• Line 141: add s to restraints.
• Line 142: add an in front of ADP.
• Line 142-143: R1 values do not fit to the ones in the table.
• Line 150: Table 3 is wrong.
• Line 151: Table 4 is wrong.
• Line 152: Figure 2 -> In the text it is not said that the XNDP keyword is used but it is said in the figure. If such keyword is used for the refinement, state this in the text and explain what this keyword means. Only intramolecular H-bond are highlighted and their distances displayed, what about the intermolecular H-bonds? Further, no table is available neither the cif file allows to confirm such values.
• Line 165: “overall structure accuracy is improved”, compared to what?
• Line 173-174: “data merging software will take radiation damage into account …”, this will be the future? Is applied here? If not, how do the authors know that such approach will still add errors into the results if still not applied?
• Line 175: accurate and reliable structure in comparison to what? Please clarify in the text.
• Line 183: XNDP keyword is a very explicit command from SHELX. Specify what this does to the ADPs.
• Line 188: Where can I find the 0.20(11) ADRA value? Specify.
• Line 198: Figure 1 -> There are 2 images in this pdf related to figure 1, which one should I look at? Further, electrostatic potential is a physical magnitude represented in Volts (V) or, commonly found in literature, e/A. e/A3 belongs to electron density related to X-ray or neutron data, but electron diffraction, as the authors stated, allows to retrieve electrostatic potentials. The authors must transform the values from densities to potentials.
• Table 1: I repeat from my last review, what authors have here are not only statistics but also structure parameters, they should state it clearer in the label caption. The authors do not need to repeat Sucrose in the title of the columns, just add RT and 100K. “Negative ADP number”, is this in A2? Please, change it to “Number of Negative ADPs” to make it clearer. The authors need to add some brackets between Fo of R1Fo. “the unit cell parameters were set to the same as these in the reference model” I understand what the authors mean but they need to rephrase it, grammatically speaking, it doesn’t make sense. Wouldn’t it be better to place the asterisk at the CC1/2[%] parameter instead of the four values? Moreover, why the authors put this when the text does not specify how important this parameter is neither refers to it elsewhere?
• Table S1: Revised the caption text, “… between sucrose the structure determined from …”. This table is not cited in any part of the text.
• Table S2: In the response to the first review, the authors say that very strong ISOR restraints were applied but it is not explained neither in the text nor the table caption and it does not specify elsewhere that still five of them are negative. This should be addressed for consistency, otherwise it looks like some details are being hidden.

H4ABTC crystals

General comments:

• This part is better structured and easier to follow than the sucrose case but it still needs a lot of improvement to properly prove the conclusion that a merged RT dataset is better or equal than a 100 K dataset for structure refinement. ADPs are not mentioned anywhere and they are not reported elsewhere in the publication. These are important values for the evaluation of the structure and the authors must include them and discuss it for fair comparison between the different temperature structures. If specific restraints/keywords are still used in here, they have to be mentioned and properly explained. Furthermore, bond distances for the 100K are not reported anywhere, thus a comparison cannot be made about the stated increased accuracy of the RT structure.
• There is a serious issue here with respect to Hydrogens. As I understand from the text, the final refinement of the RT structure was done with all placed Hydrogens (via found peaks or placed from a chemical sense) while the 100K structure was done only with the found ones, without the ones at the carboxyl group. That is not a fair comparison between the 2 temperature structures since Hs scatter strongly in electron diffraction and they may be the reason why the R1 value is a bit worse for the 100K case as well as why the disordered Os are not placed in a “good” geometrical orientation as in the RT case. H-bonds cannot be looked at because they are not available in form of a table and Hs are not included in the cif files provided to the journal to check them. This must be addressed for publication of such study since Hs are discussed quite a lot in this text.

Specific comments:

• Line 204: is this an unknown structure? It was not mentioned in the first version of the manuscript since a paper related to its synthesis was provided. If it really is, authors must explain how they synthesized the crystals for reproducibility purposes and mention the possible polymorphs already in literature.
• Line 215: “The linker” should not be there.
• Line 218: Taking into account 0.46 degrees/s, 100-110 degrees of angular range and 0.5 s/frame, results in a total of about 435-480 patterns, change it accordingly for these more realistic values.
• Line 223: If the data was merged (as it seems), it needs to be said.
• Line 225: 0.83 A is correct? Table says 0.8 A and 0.82 A. Anyway, Table 5 is wrong.
• Line 227: Ice rings are not reduced at 100 K. The resolution claim is very difficult to corroborate from these pictures, authors need to provide other strong means.
• Line 228-229: “78 unique non-hydrogen atoms” was already said before, delete it from here.
• Line 240: instead of 8143, shouldn’t it be 8936?
• Line 247: Figure 4C is referred only to the supposed disorder of the carboxyl groups. Since H-bonds distances are not discussed anywhere else in the text, they can be deleted from this image and one molecule is enough to show what the authors want to display. If H-bonds are included in a further revision of the manuscripts, they can let them here.
• Line 247-248: The discussion of the located Hydrogens is interesting since it seems like is an important part of the text (given that some more text is introduced far below). However, so less information is provided here to which are located and which are introduced from chemical reasons that this sentence is out of context.
• Line 248-249: “It is evident … structural details”, that is a conclusion sentence that has nothing to do with the context. Delete it.
• Line 249: Authors need to start a new paragraph here.
• Line 254-256: If this is true, the final refinement was done only with the found H? Why this if they can be added manually and produce a fair comparison with the RT structure? How many were found for the RT structure? Authors must clarify this.
• Line 262-264: this may be because you don’t add all Hs.
• Table 4: I repeat from my last review, what authors have here are not only statistics but also structure parameters, they should state it clearer in the label caption. The authors do not need to repeat H4ABTC in the title of the columns, just add RT and 100K. The authors need to add some brackets between Fo of R1Fo. Which unit cell parameters were used for the refinement? The authors should state it. Wouldn’t it be better to place the asterisk at the CC1/2[%] parameter instead of the four values? Moreover, why the authors put this when the text does not specify how important this parameter is neither refers to it elsewhere? The brackets for the unit cell parameters are not for the highest resolution shell, authors must specify what this means here.

Conclusions

• Line 274-276: H positions and found ones for each merged dataset are not specified. Disorder in the carboxyl group of H4ABTC at 100 K may be biased because of missing H. ADPs for H4ABTC are not provided. Due to all these reasons, such claim cannot be made.
• Line 276: Did the authors use cooling or cryo-transfer in this work? Clearly specify it in the “Materials and Methods” section.

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 3

Reviewer 2 Report

The latest changes to the manuscript are consistent with what was request in the review. 

Author Response

N/A

Reviewer 3 Report

Authors in this publication want to show that merged datasets at room temperature compared to 100 K datasets acquired in non-dedicated cryo-electron microscopes can produce equal or better structure solutions and refinements, advising to not use cryo- or cooling if the sample is stable in the TEM vacuum and under the electron beam. Such kind of work is interesting and I think that is worth for publishing.

After two heavily changes in the manuscript due to two reviewing processes, I see that the manuscript has been strongly improved that I would consider for publication after minor revisions. 

Abstract

Line 18: that should be placed in front of 3D ED/MicroED to make the sentence understandable. It was mention in the previous review, authors replied “We have made modifications accordingly.” but it was not implemented.

Results and Discussion

Sucrose

• Figure 1: The units for the threshold has been changed to Volts but the values themselves have not been changed, which doesn’t make that much sense. Authors should revise it again.
• Line 114: “nearly all atoms” would refer, let us say, to about more than 90% of the atmos. 14/23 is about 60% of the atoms, nearly all cannot be said here. Authors should change it.
• Table 1: The table label is confusing regarding the resolution used and what it is exactly in-between brackets. I understand that 0.8 Å is the resolution used for the data processing, and the brackets value refers to all reflections available in the dataset. Authors should change it accordingly otherwise it is not understandable at all and it confuses more than helps. The same for the “The resolution cut …” sentence. “The value in brackets behind … “ is redundant, you already said it before.
• Line 128: Authors should add “at” in front of “cryogenic temperature” and “room temperature”.
• Line 130: Authors should change “the atoms” for “ADPs” and “normal” for “positive”.
• Line 145: and table S1?
• Figure 2: Authors should rephrase the end of the first sentence.
• Line 169: should it be 14 atoms instead of 22?
• Line 173-174: The sentence that starts with “In addition, …” regarding the increase of mosaicity when using cooling is a bit misleading because if a sample is stable under the vacuum, such effect may not be that strong in comparison to ice contamination. I would suggest the authors to take it out or modify it to not give the impression that is critical.

H4ABTC

• Table 4: The same comment applied to table 1. In fact, the response to the comment in the previous review was better as a table caption than the current one.
• Figure 4. The sentence “Their occupancies …“ is not needed here. The other text already described what the figure contains, such statement in a figure caption is not required and it is not a main point of the publication.
• Line 213-217. From my opinion, it would be worth to mention here that no negative ADPs were found in this comparison since there were some in the case of the sucrose analysis.
• Line 226. Authors should change “normal” for “single”.
• Line 227. C-O should be C=O and C-O should be C=O.
• Line 241-244. That is quite interesting because is the inverse situation of the sucrose case and authors should comment at least with some sentences about it. Moreover, the elongation is along one specific direction in both temperature datasets, which should have a physical meaning.
• Figure 5. Authors should add “at” and “and at” in front of “(A)” and “(B)”, respectively.
• Line 260. Authors should change “normal” for a less vague word. “The expected”?
• Line 260-261: Such general sentence is misleading since ADPs are bigger for cryo temperatures in the H4ABTC case, hence the atoms are less localized and may be correct here. But in the sucrose case the ADPs are much smaller (as expected for low temperature measurements) and, in principle, fine details should be better spotted. The authors should tune this sentence in the manuscript to avoid such kind of confusion or rephrase it entirely.

Author Response

Please see the attachment

Author Response File: Author Response.docx