[Major Compulsory Revisions]

(1) The authors should explain how to handle “3D” imaging MS datasets more carefully. Page 13, line 8 “Instructions for loading the imzML files” is somewhat misleading since the description is about loading imzML, not necessarily imzML and its binary file of the “3D” imaging datasets submitted. There are a few software tools listed in the manuscript, but the authors should clearly indicate whether they are “3D” compatible or not.

Although this reviewer could manage to take a look at the m/z and intensity arrays in the binary file by writing a script (on the other hand, failed to load on BioMap and other software), prospective audience may have difficulty to figure out how the datasets look. Since 3D imaging MS is a new research field and the datasets submitted are huge amount of data, the authors are strongly encouraged to provide more useful information for prospective audience, rather than just uploading enormous amount of data.

For instance, several images of particular m/z (preferably known metabolites) from one of the sections, and several spectrum files (mzML) from particular pixels may be useful for prospective audience and developers to have an idea of the information contained in the dataset, especially when they start writing programs. Also, just like the authors prepared for 3D DESI data, providing an imzML data of “2D” image extracted 3D data as an example would be useful since it can be quickly loaded on the normal IMS viewers.

[Minor Essential Revisions]

(2) The reviewer has found numerous inconsistencies in “Data description”. For instance, some numbers do not have comma (“2,000-20,000 m/z” in kidney part, “2000-20000 m/z” with in carcinoma part). The authors must review the manuscript more carefully and standardize the writing method.

(3) The authors should explain the details of the method for 3D stack or 3D image registration. What kind of methods was used in SCILS lab ? More explanation and the citation are needed since SCILS lab software played an important role in generating 3D IMS data.

(4) P5, line9 The line “Thousands of millions of m/z values …..” is misleading. Where does “thousands to millions” come from ? Number of peaks for data points ? Although this reviewer assumes the number comes from data points of mass spectrum, it must be clarified.

(5) P5, line 23 Publications on MALDI and DESI mechanisms should be referred to.

(6) P12, line 25 Please explain how co-registration and transformation were performed. What kind of algorithm and library were used ?

(7) P14, line 11 “Potential use” is too short and must be elaborated. There the authors just briefly mentioned the general data processing methods, nothing special for 3D.

Level of interest An article of importance in its field Quality of written English Acceptable Statistical review No, the manuscript does not need to be seen by a statistician. Declaration of competing interests I declare that I have no competing interests.

Authors' response: (http://www.gigasciencejournal.com/imedia/1407920888156393_comment.pdf)

Dear editors,

The manuscript has been improved significantly, especially regarding the usability and visibility of the dataset. The presented data will be a valuable dataset for the community and facilitate development on 3D MS imaging data informatics studies. The manuscript is now suitable for publication.

Level of interest An article of importance in its field Quality of written English Acceptable Statistical review No, the manuscript does not need to be seen by a statistician. Declaration of competing interests I declare that I have no competing interests