Reviewer 1's Review (Gábor Balázsi)

-----

In this manuscript, the authors analyze a high-confidence S. cerevisiae protein interaction network and introduce the novel concept of motif date hubs (mDH) and motif party hubs (mPH). They show that mPH and their constituent motifs co localize to the nucleus, while mDH and their constituents are separately localized, apart from the nucleus. Removal of mPHs has less influence on network connectivity than the removal of mDHs. mPHs are indicative of a function associated with "nucleus", while mDHs with various other functions.

While this study required a significant amount of work, and lead to some potentially interesting results, I am reluctant in making a recommendation at this point, because the methods are not described with sufficient clarity and in sufficient detail to assure that the results are correct. The manuscript is poorly written, and difficult to read. The sections do not serve a clear role, for example the Discussion continues presenting results instead of analyzing and evaluating what was described before. The Results section contains many unnecessary details, while important details are vaguely described or omitted. The reference list is incomplete, and the Methods section is vague. The manuscript should be rewritten, necessary definitions included, and unnecessary details moved to a Supplementary Text. Towards the end, the writing becomes especially poor, at least I was unable to follow what the authors were trying to describe about the pathways.

For example, the authors refer to the earlier results by Han et al. They should describe more clearly the earlier results so that the reader does not have to re-read it to clarify what the hubs' average PCC is. Quantities such as Complex_ratio_same should be defined by formulas in the Methods, and should be defined differently (in words) and clearly in the Results, so that the reader understands the essence of this measure without reading the Methods. Results are described repeatedly, without adding new insight (the fact that mPHs and their constituents localize to the nucleus is mentioned repeatedly). All of the results sections should be reduced in size, following three questions: 1. What was the question asked? 2. What were the methods used to answer the question? 3. What was the outcome of the analysis?

The two types of hubs are essential in this manuscript, but are defined vaguely and in the wrong place in the Results (the reader has to refer to the Methods to understand what they are). mPHs and mDHs should be defined at the very beginning of the results section. Moreover, it should be clarified how they relate to the DHs and PHs defined by Han et al (is there any significant overlap? Why are they called similarly?).

The Methods section is incomplete and vague. Concepts should be very clearly described, and formulas given. Most importantly, the authors should clarify how they handle proteins that are parts of multiple complexes, or those that that belong to multiple subcellular localizations. They should tell exactly what motifs they are using (or even show them as part of a figure), not only give their numbers 238 and 13260.

The concept of "motif density" is clearly related to the "clustering coefficient", and it might not be necessary. The authors should calculate the "clustering coefficient" for nearest and next-nearest neighbors, and see how their results compare with using the "motif density". Relevant literature that should be consulted and cited includes Wuchty et al, Nat. Genetics 2003, Dobrin et al., BMC Bioinformatics 2004, Vazquez et al., PNAS 2004.

-----

N.B. These are the general comments made by the reviewer when reviewing this paper in light of which the manuscript was revised. Specific points addressed during revision of the paper are not shown.