Integrating networks with my data-naming issue

[susana.b...@nasa.gov]

Dear All,

First, thank to all that helped me last week in my first steps within
Cytoscape.

Please, let me know if what I want to do makes sense. I have
expression data mostly from qPCR, which I have entered into a small
hand-made network. I want to see how this expression data may indicate
the activation or suppression of a certain signalling pathway, so I
thought that I could import from web services the putative networks
involved and then merge my own network with those imported. Let me
remind you that my own network only has node and expression data so
far, and no edges.

I realize that Cytoscape needs a way to reconcile the names I gave to
the nodes in my network with those in the imported network. So in
order to get the "standard" names, I followed the directions provided
in the manual using the BioMart client. Which ID names should I
import? I tried ENTREZ and ENSEMBLE names, but nothing happened. Am I
going in the right direction? What am I doing wrong? If anyone can
help me with more detailed instructions, that would be wonderful.

Thank you very much!

Susana

[piet molenaar]

Hi Susana,

I think you're heading in the right direction; the way to go is usually to load the networks of interest and than load your expression data that should be coupled to an id that can be mapped to an id provided in the uploaded networks. What kind of id do you use for your data?

Piet

--
Piet Molenaar
p.mol...@amc.uva.nl
Department of Human Genetics, M1-131
Academic Medical Center
University of Amsterdam
Meibergdreef 9
1105 AZ Amsterdam
the Netherlands

tel (+31) 20-5666592
fax (+31) 20-6918626

[susana.b...@nasa.gov]

Thanks for your input!
I have been working all week with GenMAPP and I find it so much more
agreeable to me and my way of thinking. In fact, I am successfully
importing my expression data into pathways already. However, and
because I found GenMAPP so much more strightforward, I have quit for
now trying Cytoscape, so I haven't solved the problem of the IDs yet
(which was very easy to do with GenMAPP). One difference I see already
that makes Cytoscape more complete: when you plug in the expression
data, I should be able to define a color range and the range would be
assigned "continuously" and not discretely to the expression values.
GenMAPP, on the other hand, only assigns discrete colors according to
a criterion that we define (up- or downregulated within a window
value).
So far, these two programs expect us to "know" the pathway a priori,
right? Do you know of any software that allows me to just enter the
expression data and establish correlations that will allow to design
putative pathways? I have heard of and talked with the people of
Biobase. Their product seems to be able to do that but it is not free
and instead, very expensive!

Susana

> p.molen...@amc.uva.nl

> Department of Human Genetics, M1-131
> Academic Medical Center
> University of Amsterdam
> Meibergdreef 9
> 1105 AZ Amsterdam
> the Netherlands
>
> tel (+31) 20-5666592

> fax (+31) 20-6918626- Hide quoted text -
>
> - Show quoted text -

[piet molenaar]

Hi Susana,
Simplifying mapping of ids might be something we should work on (or at least the documentation; in a future version this will be easier). There are ways of integrating GenMAPP with cytoscape; probably some people from that project could jump in on this thread.
For automatically deriving networks you can consider the ExpressionCorrelation plugin; available from the plugin manager in Cytoscape. It is maintained by Gary Bader's group: http://www.baderlab.org/Software/ExpressionCorrelation
Hope this helps and if you have trouble using these don't hesitate asking!
Cheers,
Piet

--
Piet Molenaar
p.mol...@amc.uva.nl

[Alexander Pico]

Hi Susana,

I’m a developer for both the GenMAPP and Cytoscape projects. You’ll be happy to learn that we are currently working on a suite of Cytoscape plugins that bring some of the GenMAPP ease-of-use features and interfaces to the powerful Cytoscape core. We’re calling this project GenMAPP-CS and should have an alpha release out in May.  Perhaps we can include you in our early testing? It’d be great to get your feedback as a user of both tools.  BTW, effortless ID mapping will be a key feature :)

In terms of identifying relevant pathways. GenMAPP has the MAPPFinder tool which ranks pathways from a pre-existing archive based on statistical overrepresentation.  I would also refer you to Gary Bader and the makers of the Cytoscape plugin, BiNGO, for undocumented ways to identify pathways, in addition to GO terms, based on your data.

Designing new pathways based on data is yet another topic. We are very keen on making tools to help with this as well. We have an initial version of the  Bubble Router plugin available for Cytoscape (via the plugin manager). See the help link to learn more about it.  And we are in the process of making a second version of this plugin that is dramatically easier to use and more interesting.  Again, look for that around April or May.

 -Alex

[susana.b...@nasa.gov]

Alex,

Excellent! I will be happy to try the release of the early version of
GenMAPP-CS.

Thank you for all the suggestions. I have already been using
MAPPFinder (which I found useful) and am also building some
"rudimentary" pathways with MAPBuilder, as not everything is
represented in what has already been contributed online. Piet's and
your recommendation of the Bader Lab's Expression Correlation Pluggin
makes so much sense for what I need to know (which is pathway
discovery), but I ran into the following obstacle:
when I tried to import my Expression Matrix, it rejected it because I
do not have p values in it. So I tried instead to import the
expression information as an attribute table, but the Expression
Correlation pluggin expects me to import an expression MATRIX. How do
I solve this problem?

Susana

On Jan 30, 2:14 pm, Alexander Pico <ap...@gladstone.ucsf.edu> wrote:
> Hi Susana,
>
> I¹m a developer for both the GenMAPP and Cytoscape projects. You¹ll be happy
> to learn that we are currently working on a suite of Cytoscape plugins that
> bring some of the GenMAPP ease-of-use features and interfaces to the
> powerful Cytoscape core. We¹re calling this project GenMAPP-CS and should
> have an alpha release out in May.  Perhaps we can include you in our early
> testing? It¹d be great to get your feedback as a user of both tools.  BTW,
> effortless ID mapping will be a key feature :)
>
> In terms of identifying relevant pathways. GenMAPP has the MAPPFinder tool
> which ranks pathways from a pre-existing archive based on statistical
> overrepresentation.  I would also refer you to Gary Bader and the makers of
> the Cytoscape plugin, BiNGO, for undocumented ways to identify pathways, in
> addition to GO terms, based on your data.
>
> Designing new pathways based on data is yet another topic. We are very keen
> on making tools to help with this as well. We have an initial version of the
> Bubble Router plugin available for Cytoscape (via the plugin manager). See
> the help link to learn more about it.  And we are in the process of making a
> second version of this plugin that is dramatically easier to use and more
> interesting.  Again, look for that around April or May.
>
>  -Alex
>

> On 1/30/09 7:34 AM, "piet molenaar" <piet....@gmail.com> wrote:
>
>
>
> > Hi Susana,
> > Simplifying mapping of ids might be something we should work on (or at least
> > the documentation; in a future version this will be easier). There are ways of
> > integrating GenMAPP with cytoscape; probably some people from that project
> > could jump in on this thread.
> > For automatically deriving networks you can consider the ExpressionCorrelation
> > plugin; available from the plugin manager in Cytoscape. It is maintained by
> > Gary Bader's group:http://www.baderlab.org/Software/ExpressionCorrelation
> > Hope this helps and if you have trouble using these don't hesitate asking!
> > Cheers,
> > Piet
>

> >>> > - Show quoted text -- Hide quoted text -

[piet molenaar]

Hi Susana,
As I understand from the documentation you have to use an expression matrix for this specific plugin; this can contain either p-vals, or fold changes  (probably developers of the plugin can correct me if I'm wrong?).
If you manage to format your data (using excel or something like that) in this way:

GENE COMMON gal1RG gal4RG gal80R
YHR051W COX6 -0.034 0.111 -0.304
YHR124W NDT80 -0.090 0.007 -0.348
YKL181W PRS1 -0.167 -0.233 0.112
YGR072W UPF3 0.245 -0.471 0.787
YHL020C OPI1 0.174 -0.015 0.151
YGR145W YGR145W 0.387 -0.577 -0.088
YGL041C YGL041C 0.285 -0.086 0.103
YGR218W CRM1 -0.018 -0.001 -0.018
YOR202W HIS3 -0.432 -0.710 0.239
YCR005C CIT2 0.085 0.392 0.464
YER187W KHS1 0.159 0.139 -0.045
 
where each col is separated by a space and first col is id, second a common name, rest are conditions, it should be recognized as an expression matrix. It worked for me...
For specific issues you might want to contact the baderlab directly: This plugin is maintained by Laetitia Morrison and Shirley Hui in the Bader Lab, University of Toronto.
Hope this helps!
Piet

--
Piet Molenaar
p.mol...@amc.uva.nl