"PT Bote" <ptbote86 at="" googlemail.com=""> writes:
> I have been using the package Ringo to analyze some ChIP-chip data
> some months ago
[snip]
> The peaks found back then all have a slot "modification" that stored
> on what kind of ChIP sample the peaks had been found. In this new
> version, the function "newPeak" and others do not mention such a
> slot anymore and it's not used by other functions any
> longer. Unfortunately I have used this slot in some of my own
> functions and these do not work anymore. I suspect is has been
> replaced by the slot "antibody", am I right? Is there a simple way
> to update the "peak" objects?
I expect the Ringo maintainer will have some details for you...
I wanted to point out to anyone listening that any time you directly
use a slot in a class by using '@' you are very likely
going _around_ the intended interface. Many package authors use
classes and methods specifically to separate the implementation
details like slots and slot names from the user interface (the
methods). The lesson? If you find yourself using '@', ask yourself
if there is a way to get what you want without direct slot access. If
not, send a question to the list and/or maintainer.
Hope that helps (or that it will help next time) :-)
+ seth
--
Seth Falcon | Computational Biology | Fred Hutchinson Cancer Research
Center
BioC: http://bioconductor.org/
Blog: http://userprimary.net/user/
Dear Pos,
hello again. You are right, in the development version, the element
name
"modification" has been renamed to "antibody", since I think this term
is more general and eventually it's the chosen antibody that defines
which genomic regions are enriched in the ChIP. I am sorry that this
change interfered with your functions. I have changed all relevant
functions in Ringo accordingly to avoid any problems with this
renaming,
but of course the renaming could have created problems with user-
created
functions. I will try to make such changes more transparent. In the
end,
I would be happy to make the class "peak" for ChIP-enriched regions as
accessible as possible for ChIP-chip users in general. And even though
I
find S4 classes a bit too tedious, maybe I should replace the current
S3
class by a less break-prone S4 one soon. (And I know that the name
"peak" is a bit unfortunate too since it rather describes a genomic
stretch that seems to have been enriched in the ChIP. The probe-level
signal in that region, however, may show one or more actual peaks or
just a long elevated plateau.) For the moment, I am afraid, you have
to
do a manual update of the "peak" objects. Say "pl" is your list of
peak
objects, then something like
pl <- lapply(pl, function(p) {p$antibody=p$modification;
p$modification=Null; return(p)})
may work for you. Please let me know if it does not.
> On another note, the input function "readNimblegen" requires a file
> similar to the provided example file example_files.txt that
apparently
> describes the used slides. I have created such a text file by typing
it
> myself, but it would be nice to have a way to generate such a file
> automatically.
>
Thank you for the suggestion. Another user had already mentioned that
the file "SampleKey.txt" that NimbleGen sends with the pair files
could
be used for this. In the development version, the scripts directory
contains a script "convertSampleKeyTxt.R" which may give some ideas
how
to generate a targets file from it.
Best regards,
Joern
