[BioC] Pathview for global and overview maps Re: using pathview about hsa01100

Luo Weijun luo_weijun at yahoo.com
Fri Aug 16 23:26:03 CEST 2013


I don’t think pathview will do that. You will need to parse the xml manually to be able to do that.Pathview’s integrated parser works with normal pathways, where enzymes and genes are defined as rectangle nodes. Here they are edges. 
Again, for these global maps, only working with compound/metabolite nodes are fine, but not so when touching other record types.
Weijun

--------------------------------------------
On Thu, 8/15/13, Nick <edforum at gmail.com> wrote:

 Subject: Re: Pathview for global and overview maps Re: using pathview about hsa01100

 Cc: "bioconductor at r-project.org" <bioconductor at r-project.org>
 Date: Thursday, August 15, 2013, 10:11 PM

 Hi Weijun,
 It is one step away from what I want. However, I
 am wondering if, for example, I have an enzyme and would
 like to highlight its substrate and product, is it possible
 to do so? For example NME3 in pyrimidine metabolic pathway,
 I want to highlight its substrate and product in such global
 map like hsa01100.

 Thanks,
 weiwei



 On Thu, Aug 15, 2013

 wrote:

 Hi
 Weiwei,

 If you just want to the highlight the metabolite/compound
 nodes (without dealing with the enzymes etc) in such Global
 and overview maps like hsa01100, you may still use pathview
 like for individual metabolic pathways. You may see an
 example output here:


 http://pathview.r-forge.r-project.org/#fig-5



 You will notice that the round compound nodes are amplified
 for better view of the data. Below is the code I used. It
 works, but takes a long time (30 min for me on my desktop).
 The time is needed because there are so many nodes to be
 edited pixel by peixel. Option same.layer = F would speed up
 the process over 100 times, although the graph looks a
 little different as expected. HTH.


 Weijun



 library(pathview)

 sim.cpd.data=sim.mol.data(mol.type="cpd",
 nmol=3000)

 data(cpd.simtypes)

 #~30 min

 pv.out <- pathview(cpd.data = sim.cpd.data, pathway.id =
 "01100",

     species = "hsa", out.suffix =
 "sim.cpd", kegg.native = T)

 #~15 sec

 pv.out <- pathview(cpd.data = sim.cpd.data, pathway.id =
 "01100",

     species = "hsa", out.suffix =
 "sim.cpd.2layer ", kegg.native = T, same.layer =
 F)



 --------------------------------------------

 On Wed, 8/14/13, Ed <edforum at gmail.com>
 wrote:



  Subject: Re: using pathview about hsa01100



  Date: Wednesday, August 14, 2013, 3:13 PM



  Hi Weijun,

  I mean if I can highlight the chemical nodes from

  hsa01100?

  thanks,

  weiwei





  On Fri, Aug 2, 2013 at



  wrote:



  Hi Weiwei,



  I don’t think pathview can work with this diagram. It
 is

  not a real KEGG pathway graph.



  However, I notice that you can do User data mapping on
 the

  web page:



  http://www.genome.jp/kegg-bin/show_pathway?hsa01100



  Of course, you have to generate the input file with
 colors

  manually. And there will no color key etc. It may not
 look

  that neat even if you finally make it. You see, the raw

  graph already has a lot of different colors, and the
 nodes

  look too small compared to the whole graph.









  My suggestion would be have a pathway analysis done on
 your

  data using GAGE or another method, and visualize the

  perturbed pathways separately using pathview.



  Weijun







  --------------------------------------------



  On Thu, 8/1/13, Ed <edforum at gmail.com>

  wrote:







   Subject: using pathview about hsa01100



   To: "bioconductor at r-project.org"

  <bioconductor at r-project.org>,







   Date: Thursday, August 1, 2013, 5:52 PM







   Hi Weijun,



   In terms of the topmost metabolic



   pathway "hsa01100", I would like to highlight

  some



   enzymes on this pathway, however, it seems the



   "circles" on the plot are metabolites instead

  of



   enzymes. The enzymes are the "edges" on the



   plot.







   Do you know how to deal with this



   question?



   Alternatively, if there is a way to



   highlight the circles instead of edges, that's also



   fine. But is there a way to find the
 "circles"

  for



   the "edge"?







   Thanks,



   Weiwei



More information about the Bioconductor mailing list