[BioC] Annotation Tools package

[Brad Ander]

Thank you for the suggestions, Alexandre.

We indeed were placing the annotation files into the data directory of
annotationTools, now we now just use our working directory by leaving
out:

>dataDirectory <- system.file("data", package = "annotationTools")

You are also correct about the header.  We got around this by deleting
the header in the annotation file, but your suggestion to skip the
comments marked '#' will make it convenient to just use the files as
provided.  As we had it, all rows/columns were read properly.

Thanks for the help and the great tool.

Brad

--
Brad Ander, PhD
M.I.N.D. Institute
University of California at Davis
Room 2434
2805 50th Street
Sacramento, CA  95817




2009/8/12 Alexandre Kuhn <kuhnam at mail.nih.gov>:
> Hi Brad,
>
> A couple of comments on your code (see below)
>
>> -----Original Message-----
>>
>> Affy Probeset -> Entrez:
>>
>> annotationFile <- "HG-U133_Plus_2.na29.annot.csv"
>> dataDirectory <- system.file("data", package = "annotationTools")
>> annotation_HGU133Plus2 <- read.csv(paste(dataDirectory, annotationFile,
>> + sep = "/"), colClasses = "character")
>
> I assume here that you have the annotation file
> "HG-U133_Plus_2.na29.annot.csv" in the 'data' subdirectory of the
> annotationTools installation directory. I used this code in the vignette to
> load an example annotation file (stored in 'data'). You can however save the
> annotation file anywhere on your file system.
>
> Second I am not sure your annotation file loaded correctly since Affymetrix
> now has a header in annotation files. To skip the header (lines preceded by
> the dash sign), please use the command (assuming for instance that you work
> under Windows and you saved you annotation file under "C:/Annotations")
>
>>annotation_HGU133Plus2 <-
> read.csv("C:/Annotations/HG-U133_Plus_2.na29.annot.csv ", colClasses =
> "character", comment.char='#')
>
> You can check the size of annotation_HGU133Plus2 with
>
>>dim(annotation_HGU133Plus2)
>
> to make sure that you now have what you expected (that is, a data.frame of
> 54675 rows and 41 columns).
>
> I changed the vignette accordingly some time ago and the change will be
> incorporated in the next Bioconductor release.
>
>
> Alexandre
>
>
>
>
>> allPS<-annotation_HGU133Plus2[,1]
>> getANNOTATION(allPS, annotation_HGU133Plus2, diagnose = FALSE,
>> identifierCol = 1, annotationCol = 19, noAnnotationSymbol = NA,
>> noAnnotationProvidedSymbol = "---", sep = " /// ")
>> entrez <- getANNOTATION(allps, annotation_HGU133Plus2, diagnose =
>> FALSE, identifierCol = 1, annotationCol = 19)
>> write.matrix(entrez, file = "humanentrez.csv", sep = " ")
>>
>> Entrez -> illumina:
>>
>> annotationFileIll <- "HumanRef-8_V3_0_R2_11282963_Ab.csv"
>> dataDirectory <- system.file("data", package = "annotationTools")
>> annotation_Illumina <- read.csv(paste(dataDirectory, annotationFileIll,
>> + sep = "/"), colClasses = "character")
>> getANNOTATION(entrez, annotation_Illumina, diagnose = FALSE,
>> identifierCol = 9, annotationCol = 14, noAnnotationSymbol = NA,
>> noAnnotationProvidedSymbol = "---", sep = " /// ")
>> illuminaID <- getANNOTATION(entrez, annotation_Illumina, diagnose =
>> FALSE, identifierCol = 9, annotationCol = 14)
>> write.matrix(IlluminaID, file = "illuminaID.csv", sep = " ")
>>
>>
>> It may not have been the most perfect use of the code but it seems to
>> work (we are still learning).
>>
>> Thanks for your help.  If there are any suggestions you feel are
>> important, please let us know.
>>
>> Kind regards,
>> Brad
>>
>>
>> --
>> Brad Ander, PhD
>> M.I.N.D. Institute
>> University of California at Davis
>> Room 2434
>> 2805 50th Street
>> Sacramento, CA  95817
>>
>>
>>
>> 2009/8/10 Alexandre Kuhn <alexandre.kuhn at epfl.ch>
>> >
>> > Hi Yingfang,
>> > Once you have loaded your Affymetrix annotation into R (assume it is
>> > contained in an R object named 'annot') you could for instance select
>> all
>> > probe sets by subsetting the data.frame so as to select the first
>> column
>> >
>> > > allps<-annot[,1]
>> >
>> > I am not sure this answers your question tough. Could you please send
>> some
>> > lines of code to help me understand what is going wrong?
>> >
>> > Best, Alexandre
>> >
>> > -----Original Message-----
>> > From: bioconductor-bounces at stat.math.ethz.ch
>> > [mailto:bioconductor-bounces at stat.math.ethz.ch] On Behalf Of Yingfang
>> Tian
>> > Sent: jeudi 6 août 2009 19:23
>> > To: bioconductor at stat.math.ethz.ch
>> > Cc: Brad Ander
>> > Subject: Re: [BioC] Annotation Tools package
>> >
>> > Dear Dr Kuhn:
>> >
>> > We are Yingfang Tian and Brad Ander from the University of California
>> at
>> > Davis.  We are working on the cross-platform analysis of 3 platforms:
>> > llumina Human Ref-8,  Affymetrix U133 plus 2 array, and Affymetrix
>> human
>> > Exon array.
>> >
>> > We are trying to use your annotationTools package in R amd are able
>> to at
>> > least translate across probes from U133 arrays to Illumina, similar
>> to the
>> > example you give in the BMC Bioinformatics paper.   We were wondering
>> how to
>> > import large or entire numbers of probesets into the “myPS” object?
>>  It may
>> > be a basic R command, but unfortunately we rely on commercial
>> software for
>> > the majority of our analyses and have limited experience with R
>> (hopefully
>> > that can change on both fronts).
>> >
>> > >From the paper, it seems that we can generate a list of Refseq IDs
>> from
>> > >the
>> > Affy Probesets and then use this list (set it as the “myPS” object)
>> to pull
>> > out the Illumina Probe ID using the refseq column as the identifier
>> column.
>> > We can export all these with a simple write command.  Right now, we
>> are
>> > thinking of bridging to/from the Exon arrays with the Unigene.  I
>> guess we
>> > will have to see how that works
>> >
>> > Again, we are having success when dealing with a few probesets, but
>> there
>> > must be a way to get ALL probesets.  Can you please help us with
>> this?
>> >  Possibly
>> > with the example command syntax?  In the paper you mention mapping
>> all the
>> > mouse probes across platforms, so you must have had to deal with
>> this.
>> >
>> > We are likely wanting to try the cross species analysis in the near
>> future
>> > as well, so learning how to get passed the limit of entering each
>> > probe/gene/etc manually will be a big help.
>> >
>> > Kind regards,
>> >
>> > Yingfang and Brad
>> >
>> > --
>> > Yingfang Tian, PhD
>> > M.I.N.D. Institute
>> > University of California at Davis
>> > 2805 50th Street,Room 2434
>> > Sacramento, CA  95817
>> > Tel:916-703-0384
>> >
>> >        [[alternative HTML version deleted]]
>> >
>> >
>
>