[BioC] tilingArray getting started questions

Wed Sep 8 04:19:30 CEST 2010

Dear Wolfgang,

Thank you for the helpful response and for creating with your colleagues this excellent tool.  

1.  I will let you know if I encounter any problems with passing ExpressionSet objects from Starr into the normalizeByReference function.

2.  I have only worked with ChIP dsDNA products and tiling arrays so the hybridizing of single strand cDNA would seem (to me) to bring a different set of technical concerns.  I have read the paper describing the Actinomycin D use during reverse transcription and have included this step in my experiments.  It was probably not the appropriate forum to probe for experimental tips.

Best,

noah

On Sep 7, 2010, at 12:49 PM, Wolfgang Huber wrote:

> Dear Noah
> 
> On 07/09/10 18:34, Noah Dowell wrote:
>> Hello All,
>> 
>> I would like to use the tilingArray package to analyze transcript
>> products from several S. cerevisiae mutants using the Affymetrix
>> Tiling Array 1.0R.  (Note that these arrays only have one strand of
>> the yeast genome tiled on the array.)  I have used these tiling
>> arrays for ChIP-chip experiments.  To analyze that data the Starr
>> package was extremely helpful.  I have reproduced the examples
>> (using the davidTiling data) in a couple of the tilingArray vignettes
>> and I am getting ready to move into using my data, but have several
>> questions that do not appear in the vignettes or on the archives.
>> The general theme of my questions is if anyone has any experience on
>> the compatibility between Starr and tilingArray.
>> 
>> 1. The Starr package has a function (bpmapToProbeAnno) to create a
>> probeAnno object (mapping of the array features to the yeast genome).
>> Can I use this function to create a probeAnno for tilingArray?
> 
> The answer depends on which function in the 'tilingArray' package you refer to. AfaIcs, there are for main ones:
> 1. normalizeByReference - does not use probeAnno
> 2. segment - does not use probeAnno
> 3. segChrom - a wrapper around 'segment' that does some bookkeeping and looping, and which is intended (and documented) to be able to use Ringo's probeAnno class (which is also what Starr uses). Let us know if there is any trouble with that.
> 4. plotAlongChrom - a fairly idiosyncratic function written for our particular array, and which I do not intend to work as a general purpose tool for other arrays, but components (code snippets) of which you might (or might not) find useful for creating your own displays.
> 
> Hence, the answer is
> [1]   NA   NA   TRUE    FALSE
> 
>> 2.  Reading my CEL files in using Starr (function: readCelFile) can
>> easily create an ExpressionSet object.  Should/can I pass this eset
>> to the tilingArray normalizeByReference function after making sure to
>> point to the correct RNA and DNA samples?
> 
> Yes.
> Let me know if there is any trouble with that.
> 
>> 3.  Alternatively, I could normalize the arrays independently in
>> Starr (function: normalize.Probes (method="loess") and then use the
>> Starr getRatio function to normalize to the "reference" genomic DNA
>> samples.  The output of the getRatio function will be an eset that I
>> think I could pass directly to the segChrom function.  Am I thinking
>> correctly about this?
> 
> Yes.
> 
>> 4.  In David et al, a custom tiling array from Affymetrix was used
>> that had probes corresponding to both strands of DNA.  Since I have
>> amplified my single-stranded cDNA with Taq to make a double stranded
>> DNA product which is then hybridized to the array it seems like the
>> tiling arrays that only have one strand (the commercial ones) should
>> work fine for mapping transcripts from both strands.  Am I missing
>> something here in the experimental workflow?
> 
> This sounds reasonable.
> 
>> If one were to hybridize
>> cDNA directly to an array it follows that both strands of the genome
>> should be represented on the array, but wouldn't lower limit of
>> sensitivity issues then be of concern?
> 
> I am not sure I understand what the concern is. This is what we did in a series of papers (see e.g. PMID 20193063, 19536197, 19169243, 16569694). It turned out to be crucial to use actinomycin D (ActD) during reverse transcription (PMID 17897965)
> 
>> I understand if people have not tried to tie the Starr/Ringo and
>> tilingArray packages together, but I thought I would ask a few
>> questions before spending significant time working out the kinks.
>> Thank you in advance for any help!!
> 
> Thanks,
>  Wolfgang
> 
>> 
>> Best,
>> 
>> Noah
>> 
>> 
>>> sessionInfo()
>> R version 2.11.0 (2010-04-22) i386-apple-darwin9.8.0
>> 
>> locale: [1] en_US.UTF-8/en_US.UTF-8/C/C/en_US.UTF-8/en_US.UTF-8
>> 
>> attached base packages: [1] grid      stats     graphics  grDevices
>> utils     datasets  methods   base
>> 
>> other attached packages: [1] geneplotter_1.26.0   lattice_0.18-5
>> annotate_1.26.0      RColorBrewer_1.0-2   davidTiling_1.2.12 [6]
>> GO.db_2.4.1          RSQLite_0.8-4        DBI_0.2-5
>> AnnotationDbi_1.10.0 tilingArray_1.26.0 [11] pixmap_0.4-10
>> Biobase_2.8.0
>> 
>> loaded via a namespace (and not attached): [1] affy_1.26.0
>> affyio_1.16.0         genefilter_1.30.0     limma_3.4.0 [5]
>> preprocessCore_1.10.0 splines_2.11.0        strucchange_1.4-1
>> survival_2.35-8 [9] tools_2.11.0          vsn_3.16.0
>> xtable_1.5-6 _______________________________________________
>> 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
> 
> -- 
> 
> 
> Wolfgang Huber
> EMBL
> http://www.embl.de/research/units/genome_biology/huber
> 
> 