Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

Adding to GoSurfer website RE: A couple comments on http://bioinformatics.oxfordjournals.org/cgi/reprint/bti565v1

Sheng Zhong zhong at Stanford.EDU
Fri Aug 5 15:32:35 PDT 2005


Sorin,

Thank you for the note. I agree that my knowledge on highly relevant papers
like yours is very limited. That is partly the reason I found your survey
paper interesting :)

Best regards,
Sheng 

********************
Sheng Zhong
Assistant Professor
 
Dept of Bioengineering & Dept of Statistics
University of Illinois at Urbana-Champaign
http://bioinfor.bioen.uiuc.edu
********************
 
 
________________________________________
From: Sorin Draghici [mailto:sorin at wayne.edu] 
Sent: Friday, August 05, 2005 2:57 PM
To: Sheng Zhong
Cc: 'Barry Zeeberg'; go-friends at genome.stanford.edu;
claude.pasquier at gmail.com; 'John N. Weinstein'
Subject: Re: Adding to GoSurfer website RE: A couple comments on
http://bioinformatics.oxfordjournals.org/cgi/reprint/bti565v1

Sheng, 

Thanks for your email. Sure, feel free to add the paper on the GoSurfer
website. I will make sure I reference the papers you indicated in any future
publications. It would be nice if you could also refer to our previous work
on this topic. I noticed for example that the recent paper that you just
sent me, does not refer to any of our 6 papers on this topic published
between 2001 and 2004. :-) 

I will contact you separately regarding the book. An analysis of the
difficulties of corrections for multiple comparisons in the GO context would
be a very valuable addition. 

Regards, 

Sorin 


Sheng Zhong wrote: 
Dear Sorin and Barry,

Thank you for the interesting discussion. 

Sorin: I have found your paper very interesting and refreshing to read. If
you allow, I would like to add it onto GoSurfer's website as a reference.

I like your point about the heavily correlated GO structure that complicates
the multiple hypothesis testing issue. Because of this issue, no standard
FDR calculation is valid. We also had an article that extensively discussed
this issue and proposed a treatment that makes valid FDR estimate for every
GO term. Here is a link to this paper:
http://www.stanford.edu/group/wonglab/doc/Zhong_S_GoSurfer2.pdf

I am happy to contribute any materials you need for the book chapter.

Best regards,
Sheng

********************
Sheng Zhong
Assistant Professor
 
Dept of Bioengineering & Dept of Statistics
http://bioinfor.bioen.uiuc.edu
********************
 

-----Original Message-----
From: owner-gofriends at genome.stanford.edu
[mailto:owner-gofriends at genome.stanford.edu] On Behalf Of Sorin Draghici
Sent: Wednesday, August 03, 2005 3:49 PM
To: Barry Zeeberg
Cc: go-friends at genome.stanford.edu; claude.pasquier at gmail.com; John N.
Weinstein
Subject: Re: A couple comments on
http://bioinformatics.oxfordjournals.org/cgi/reprint/bti565v1

Barry,

Thank you for your email. The paper dicusses the tools as they were 8-9 
months ago, when we did this work. As you know, publishing paper 
involves a certain time lag: submission, review, address any comments, 
etc. The publisher does not allow further modifications of the paper at 
this time but we have noted your comments and corrections and I will 
include them in any future paper on this topic. For the time being, I am 
currently working on a book on this topic to be published by Chapman and 
Hall/CRC Press. The book is focused on tools for ontological analysis. I 
would use this opportunity to inviite you and your collaborators (John 
Weinstein, etc) to contribute a chapter about GO Miner and other related 
tools you might have developed in this area. This would give you a first 
hand opportunity to present the latest and greatest capabilities of your 
tools. If you would like to contribute with this chapter, please let me 
know and I will give you more details about it.

Best regards,

Sorin


Barry Zeeberg wrote:

  
Hi Sorin,

I just now had a chance to glance through your review, and I noticed a few
items related to GoMiner that should be corrected in an erratum. This is
    
not
  
exhaustive, as I have only had about one minute to peruse the literature.

I will first note that your paper represents an important and ambitious
attempt to provide orientation to users in this field. It is inevitable
    
that
  
some inaccuracies will creep in. Perhaps you could host  a web  site where
authors of the cited papers can correct inaccuracies - maybe even come out
with a second edition of the paper that contains all of the corrections?

(1) GoMiner is a java application, and is therefore platform-independent,
not restricted to Windows only. I have run it on mac os x (Unix) and
    
Solaris
  
a _lot_ of times.

(2) The timing comparison should be done  using a local  database, since
hitting on our database from a remote site will depend on network  traffic
etc. I did not read carefully enough to know if you did correctly use a
local database that was properly  installed. The presentation was a little
ambiguous, in that GoMiner requires a total and a changed gene file as
input. It is not clear if the number  of genes on the x axis refers to the
total or to the changed gene file? In fact, you only go up to 1000 genes
    
max
  
- this is silly, since we routinely analyze 10,000 - 20,000 genes in the
total gene file (representing all  of the genes on  a microarray). Assuming
a changed genes file of around 400 or 500 genes, and using a local
    
database,
  
the analysis is done in around the 350 seconds that your graph implies for
1000 genes.

(3) GoMiner supports all GO identifiers, as well as HUGO names. HUGO names
are not supported by GO database, but we have a freely-available
    
enhancement
  
program to post modify the native GO database.

(4) Your statement about Onto-Express being the only application that
collapses (assigning descendant  genes to parent) is wrong. GoMiner does
this also.

(5) You state that GoMiner uses a local database server. The context of the
statement strongly suggests that there is no alternative. In fact, GoMiner
can use either a local database server or our database server, not to
mention any other accessible build of the GO database hosted anywhere.

[WARNING: the following contains  some philosophizing and self-serving
    
stuff
  
that is not mandatory reading,and the casual reader might want to skip]

You might have considered a comparison of the ability of the user to
hyperlink out from the tool to find more information about the genes or
categories of interest. For instance, GoMiner allows access to biocarta and
kegg pathway maps, genecards, NCBI Entrez, NCBI 3D structures, etc. There
    
is
  
always a danger in a review of this type for the author to focus on
    
features
  
in which her or his tool excel. This is particularly true since each person
will have a somewhat different view of what is important, and this
    
viewpoint
  
will  coincide perfectly only for that author doing the review and her or
his own tool. I would certainly not claim for myself sufficient objectivity
to value a  different  point of view to the same extent that I value my
    
own.
  
As an admittedly self-serving bit of advertising for which I apologize in
advance, I  will mention that there is a recent  paper on High-Throughput
GoMiner
( http://discover.nci.nih.gov/gominer/htgm.jsp ).
This is a web based  application and therefore is platform independent.
Among other things, it permits _integrative_ analysis of multiple
    
microarray
  
(or any other type of high throughput experiment), it implements a multiple
comparisons correction, it provides visual information on the transcription
factor binding  sites in the genes and categories, and it provides
    
clustered
  
image map (CIM) visualizations of the relationships between interesting
categories, genes, and experiments.

Finally, since the time that GoMiner was evaluated for the review article,
it has been enhanced enormously by virtue of being a pilot project in the
caBIG initiative. In  particular,  it is fully integrated at the front end
with data deposited in  caWorkbench,  and at the back end with VennMaster.

Regards,
Barry


 

    

  


-- 
Sorin Draghici, Ph.D. 


Director of the Bioinformatics Core, Karmanos Cancer Institute

Associate Professor		Tel: (313) 577-5484
Dept. of Computer Science	Fax: (313) 577-6868
Wayne State University
5143 Cass Ave, Room 431 State Hall, 
Detroit, MI, 48202
WWW: http://vortex.cs.wayne.edu/Sorin/ (personal)
WWW: http://vortex.cs.wayne.edu/Projects.html (lab)


Check out my recent book: Data Analysis Tools for Microarrays:
http://www.crcpress.com/shopping_cart/products/product_detail.asp?sku=C3154&
parent_id=&pc=



--
This message is from the GOFriends moderated mailing list.  A list of public
announcements and discussion of the Gene Ontology (GO) project.
Problems with the list?           E-mail: owner-gofriends at geneontology.org
Subscribing   send   "subscribe"   to   gofriends-request at geneontology.org
Unsubscribing send   "unsubscribe"  to  gofriends-request at geneontology.org
Web:          http://www.geneontology.org/



More information about the go-friends mailing list