Subject: Review: C582: The Virtual Laboratory: Enabling Molecular Mod From: Omer F Rana Date: Sat, 5 Jan 2002 19:15:59 +0000 (GMT) To: gcf@grids.ucs.indiana.edu CC: ccpe@grids.ucs.indiana.edu Hi, Review of C582 attached. regards Omer ---------------------------------------------------------------- C: Paper and Referee Metadata Paper Number Cnnn: ------------------ C582 Date: ----- January 5, 2001 Paper Title: ------------ The Virtual Laboratory: Enabling Molecular Modeling for Drug Design on the World Wide Grid Author(s): ---------- Buyya, Branson, Giddy and Abramson Referee: -------- Omer F. Rana Address: -------- Department of Computer Science Cardiff University 5 The Parade, Cardiff CF24 3XF, UK email: o.f.rana@cs.cf.ac.uk RECOMMENDATION: --------------- accepted provided (minor) changes suggested are made D: Referee Comments (For Editor Only) ---------------- This is a good and interesting paper -- which presents an interesting and important application for Grid Computing. So, I highly recommend this for the journal. E: Referee Comments (For Author and Editor) ---------------- This is a good and interesting paper. I hope the authors can use the comments below to improve their manuscript: It is unclear after reading the paper whether the DOCK program is suitable or not for utilisation over Grids. Was this program modified in any way -- for instance to utilise it with Globus -- or was this software used as is? Although not related to the paper directly, it is not clear how useful the spherical modelling approach adopted by DOCK actually is. How useful is modelling of complex geometries using the sphere model? It would be useful to perhaps mention how widely used this particular code is in the scientific/bioinformatics community. It is not clear from reading section 2, and from figure 2, what are the costs for initialisation. You recommend that data staging is not necessary but recommended -- however, no cost of this has been subsequently mentioned in the simulations. It is mentioned in section 2.3 that the database was only required to be replicated on a few nodes -- how is this achieved? You mention in a subsequent section (paragraph 1, page 10) that the "get_molecule_fetch" is used to retrieve data -- is this part of this process? I was unable to make a connection between the statement mentioned in section 2.3 and subsequent treatment of this issue. >From the sequence diagram on page 7, and mention of the scripts and input files on page 8, it is unclear what the user involvement is. Does the user create these files and then everything else is automatic? At what point is cost assigned? Is this part of the script? Can the cost be modified during analysis (i.e can the user interactively change a sweep range?). It would be useful, perhaps, to extend the sequence diagram on page 7 to show a complete run (which also includes the assignment of cost values). I found that the results were not discussed in detail -- for instance -- the graphs presented on page 14 are correlated in some way? The description on page 13 is very limited, and does not show or emphasise specific features in the trends. It may be useful to describe the graphs in more detail. Also, the graphs are quite dense -- i.e. there is a lot of data displayed -- it may be useful to also have some results in a tabular format (perhaps to emphasise some points) in addition to the graphs. F: Presentation Changes -------------------- Inconsistent use of spellings and caps: "Grid" or "grid" -- it may be useful to use Grid as "grid" has been used for something else in figure 4. "modelling" or "modeling" -- use one spelling throughout. Use of terms such as "millions of molecules" or "millions of ligands" etc -- is a generalisation. The authors may wish to re-phrase this to something like "a very large number (of the order of a million)" etc. On page 2, last paragraph, line 1 replace "application" with "applications" On page 2, last paragraph, line 4 replace "computational" with "computationally" On page 3, section 2, bullet point 4, sub-point 5 "selecting suitable"? -- incomplete sentence In the references section: Reference 1: Incorrect name -- replace "Future" with "New" Reference 15: Incorrect spelling for "Computational Chemistry"