- It would be useful to add a paragraph to the Discussion highlighting (1) how this approach could be utilized to better target a set of prognostic markers from patient samples and (2) the potential generalization of this approach to other cancers.
- To avoid ambiguity, it would be better to use "AUROC" or "ROCAUC" instead of just "AUC". Moreover, this abbreviation and "AUPR " should be defined at their first usage (line 185). “ROC-PR” should be “AUPR.”
- TCGA LUAD data should be cited according to guidelines given by the Broad Institute.
- For a network visualization, the networkD3 package in R is very useful.
- To improve readability of the preprint, it would be helpful to (1) print the supplementary table numbers and/or captions directly above their corresponding table contents, and (2) print the figure numbers on the same page as each figure (pages 20-30).
- Depending on the style guide, Ref. 37 should state that it is a PhD dissertation. It would also be helpful to refer to specific chapters or even sections.
- NetRank should have a capitalized R throughout.
- The authors should clarify how the EMT features were binarized using their means (line 280).
- The authors should clarify which steps were performed elsewhere and which were performed for this manuscript. For example, they previously “tested the EMT signatures” (line 344).
- Minor editing and proofreading by a generalist reader would be helpful. Some examples are listed below:
- Line 4 typo: caner --> cancer
- Line 9 typo: details --> detail
- Line 132 typo: being --> are
- Line 137 typo: should use a semicolon, as in “samples; we thus”
- Line 139: should be 74, 455, 123 nodes, respectively
- Lines 206-207: “association rule mining” can be cited earlier than line 269. Some of the details of this approach are currently found in the Results section; they may be more appropriate in the Experiments and/or Introduction sections.
- Line 282: textit should be preceded by \ in the (presumably LaTeX) source code
- Lines 288-304: the interpretation of these rules would benefit from citations supporting the “established findings in cancer research”