Referee 2's Review (Joshua Rappoport):

Review of the original submission:

This manuscript uses an innovative strategy to detect 5 different proteins of interest in cells through fluorescent imaging. Furthermore, the authors apply sophisticated data analysis regimes in an attempt to elucidate the proteins in structures such as focal adhesions. Finally, these approaches are used to elucidate cellular roles of Rho Kinase.
The authors have clear technical and quantitative ability and the paper was tantalizingly interesting. However, there are a number of serious issues (listed below) which preclude this authors acceptance of the manuscript in the current form. If satisfactorily addressed (either though further work, or rebuttal), I would be quite happy to re-assess this manuscript.

1. How do they know they are imaging things only in the adherent membrane? This is especially true when they are imaging at cell edges where the dorsal plasma membrane would be essentially indistinguishable from the adherent with their imaging techniques?
2. Where are the controls for emission bleedthrough? It is absolutely imperative that they show single colour controls demonstrating that the filters they are using don't pick up emission of fluorophores other than that which they intend image. (see point 7, below)
3. The authors have chosen to do their analysis on a pixel-by-pixel basis. In my opinion this is problematic for several reasons:
a. How were the images aligned? Although the cells are fixed, simply moving a filter cube can cause a small alteration in XY. Were brightfield reference images taken with each filter set and compared for alignment purposes? Considering the quantification is on a pixel-by-pixel and not structure-by-structure basis, this is a very important point. Even mis-alignment by a single pixel in some of their images could create a large degree of error.
b. There do not seem to be any statistics at all? There are no error bars or p-values in any of the quantifications.
c. They make reference to specific types of structures (e.g. Focal Adhesions); however, they never define exactly how these determinations are made. Would it not be better to (1) determine specific criteria for identification of different structures, (2) quantify the average fluorescence per unit area in numerous examples of each relevant structure in several cells, and (3) do statistical analysis on these sets (e.g. means +/- s.e.m.)?
i. The Introduction states the following: "Based upon morphological and molecular criteria, several types of cell-matrix adhesions were distinguished..." What are these criteria? How were they applied? Have they been shown to be valid in this cell line under the current culture, labelling and imaging conditions?
4. Is the GFP-Beta3-integrin functional? Do it specifically localize as the endogenous? Does it form proper hetero-dimers?
5. They use the term "positively labelled" - How was this defined? What kind of threshold? 5% above background? 50%? Was this the same for all markers?
6. Is there a biochemical "Gold standard" for the chosen use of Y-27632 in these cells? Why/how was 100 microM for 3 hours chosen?
7. What exactly does "minimal cross-interference between the different antibodies" (page 12) mean?

Minor points
1. On the bottom of page 6 there is a sentence that begins, "Big part of these signatures..." - Is this proper English?

Review of the first revised manuscript:

General comment: The authors have made great strides with the current version of this manuscript, as well as the responses in the rebuttal letter. Most of the points raised in the initial review have been satisfactorily dealt with. However, there are still a few issues that need clarification (In particular, Comments 3b, 3c/d, and 7).

Review of the second revised manuscript:

Well done! The authors have definitely satisfied the concerns of this reviewer, and I feel as though the manuscript is ready for publication as is. However, I would still prefer if the authors would either show as a supplement representative images or a table with the controls for antibody cross-reactivity as well as signal resulting from incubation in secondary antibodies alone. I appreciate that this level of control is not generally expected, but most papers do not have 5 different labels in the same sample analysed on a pixel-by-pixel basis!

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.