[BioC] Problem with customised endnode list in ontoCompare(goToolspackage)

[Paquet, Agnes]

Hi Manasa,
 

The bug in CustomEndNodeList should be fixed in the new release version of goTools. There was a NA value in the list at rank=2 which caused the error, the new version will remove NAs automatically.

 

For the 3 top nodes: these nodes are included by default to ensure that the loop in the function will stop if you input some GO ids that are not children of your end nodes. Any suggestion on how to improve this function will be welcome.

 

GoTools does not add any ontology to the results, unless some of the oligos ids provided in input do not match any GO ids. In this case, these oligo ids will be added under a "NotFound" category. For more details about the "unknown" category, you should take a look at the GO graph directly using a browser like AmiGO for example.

 

Regards,

 

Agnes

 

________________________________

From: bioconductor-bounces at stat.math.ethz.ch on behalf of Manasa Ramakrishna
Sent: Thu 4/19/2007 9:14 PM
To: bioconductor at stat.math.ethz.ch
Subject: [BioC] Problem with customised endnode list in ontoCompare(goToolspackage)



Hi all,

I am using ontoCompare to produce graphics for some experiments. The
datasets

are Agilent microarray based. So i have mapped all of the IDs to GO terms
and i

feed GO terms from differentially expressed lists as an input into goTools.
The

aim then is to see how these fall into higher GO categories and draw a plot.


Often, it is a comparison of upregulated vs downregulated genelists for any

given experiment.



The scientists that i work with often have a few functional categories that

they would like their differentially expressed genes to turn up in and this
was

one of the reasons i started to use goTools as it had an option for passing
in

customised endnode lists. But so far, i haven't had this bit work
successfully

on my data.

Version: I am using R 2.4.1 and goTools 1.6.0 method = 'TIDS'

1. I have used the MFnode <- CustonEndNodeList("GO:0003674", rank = 2)

declaration and though this works fine for rank = 1, 2 this is the error

message i get for rank = 3 and anything above that. Why is that ?

> BPnode<-CustomEndNodeList("GO:0008150",rank=3)

[1] "rank= 1"

[1] "rank= 2"

[1] "rank= 3"

Error in get(x, envir, mode, inherits) : variable "NA" was not found

Error in Ontology(get(id, env = GOTERM)) :

error in evaluating the argument 'object' in selecting a method for

function 'Ontology'

2. The other style i have use is to pass these nodes directly into 'endnode'


using endnode = c("GO:XXXXXX1","GO:XXXXXX2","GO:XXXXXX3",GO:XXXXXX4") and so


on. This seems to work, except that

(a) the final categories that are plotted are the top level categories among


those included in the endnode list. So say i had "immune response",
"response

to oxidative stress", "response to chemical stimulus" and "metabolic
process"

as my endnodes, my graph displays "biological_process", "metablolism"

and "unknown".

(b) I dont fully understand what is included in the "unknown" bars of the
plot.

Could someone please elaborate on this ?

(c) The last thing i found with sending in a user-defined endnode list is
that

the top ontology(ies) - BP, CC and MF are included by default. Sometimes, we


already know that we are looking at a specific ontology and there is no need


for it to be displayed on the barchart. I see that 3 top ontology GO terms
are

part of the code in ontoCompare and i tried re-defining this but was

unsuccessful.

Any ideas, help, suggestions would be greatly appreciated.

Regards,

Manasa Ramakrishna

Computational Biologist, Agresearch Limited, New Zealand

        [[alternative HTML version deleted]]

_______________________________________________
Bioconductor mailing list
Bioconductor at stat.math.ethz.ch
https://stat.ethz.ch/mailman/listinfo/bioconductor
Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor