[BioC] FW: Using HTqPCR for Fluidigm BioMark input data (48.48 array)

Mon Nov 19 16:00:26 CET 2012

Just following my own advice here.

On 11/19/12 9:22 AM, "Fletez-Brant, Christopher (NIH/VRC) [C]"
<christopher.fletez-brant at nih.gov> wrote:

>Hi Jens,
>
>Sorry for taking so long to get back to you - I was out of town.  A couple
>of things:  
>
>1) Please always use 'reply to all', as we use the Bioconductor mailing
>list as a kind of knowledge repository for others.
>
>2) Fluidigm numbering and well number aligns in multiples of 12.  That is,
>Well01 = A01, Well13 = B01 etc.
>
>3) For calculating sample means, I have been doing that manually and then
>creating a new qPCRset.  Manual calculation has involved removing subject
>ID information from sample Ids (I.e. The row named "BJ_fibroblast_well01"
>becomes "BJ_fibroblast"), taking sample specific subsets and calculating
>gene means for each of them.  That looks something like this:
>
>samples <- gsub("BJ_fibroblast|RiPSC.HUF1", "\\1", sampleNames(myqPCRset))
> #remove sample-specific ID components
>myqPCRset.names <- unique(samples)    #get unique sample names; equivalent
>to getting the number of sample types
>inds <- lapply(myqPCRset.names, grep, x=sampleNames(myqPCRset))     #split
>into groups
>groups.myqPCRset <- lapply(inds, function(i) exprs(myqPCRset)[,i])
>myqPCRset.means <- matrix(ncol=length(groups.myqPCRset[[1]]),
>nrow=length(groups.myqPCRset))    #data structure to save means in
>for (i in 1:length(groups.myqPCRset)){
>    if (length(inds[[i]]) > 1){    #only calculate the mean if there's
>more than one in the group
>      myqPCRset.means[i,] <- apply(groups.myqPCRset[[i]], 1, mean)
>    }
>    else {
>      myqPCRset.means[i,] <- groups.myqPCRset[[i]]
>    }
>}
>
>
>I hope this is sufficient to get you started.  Feel free to ask more
>questions as needed.
>
>Kipper
>
>
>________________________________________
>From: Jens Durruthy-Durruthy [jensdd at stanford.edu]
>Sent: Tuesday, November 13, 2012 4:03 PM
>To: Fletez-Brant, Christopher (NIH/VRC) [C]
>Subject: Re: [BioC] Using HTqPCR for Fluidigm BioMark input data (48.48
>array)
>
>Hi Kipper,
>
>Thanks for your email and your help! I really appreciate it. I think I
>have a fundamental problem in understanding how the data in my Fluidigm
>.csv file are read into the qPCRset. I do see that all the features where
>recognized in the right order.
>
>> featureNames(raw)
> [1] "HSP90AB1" "CD13"     "COL1A1"   "PDGF3B"   "CD90"     "VIM"
> [7] "BUB1"     "CCNF"     "CDC20"    "CDKN1A"   "LATS2"    "MAD2L1"
>[13] "RBL1"     "BPTF"     "CBX7"     "DNMT1"    "EED"      "GLP"
>[19] "G9A"      "P300"     "EZH2"     "INO80C"   "JARID2"   "KDM3B"
>[25] "MBD3"     "MCRS1"    "MLL2"     "RING1B"   "BRG1"     "SNF2H"
>[31] "HP1"      "TAF1"     "TET1"     "THAP11"   "WDR5"     "PRMT5"
>[37] "CDH1"     "CDKN2A"   "GRB2"     "LEFTY1"   "LEFTY2"   "LMNB1"
>[43] "MAPK1"    "MAPK3"    "P53"      "WNT1"     "PIK3CG"
>
>
>But when I type in
>
>> sampleNames(raw)
> [1] "Sample1"  "Sample2"  "Sample3"  "Sample4"  "Sample5"  "Sample6"
> [7] "Sample7"  "Sample8"  "Sample9"  "Sample10" "Sample11" "Sample12"
>[13] "Sample13" "Sample14" "Sample15" "Sample16" "Sample17" "Sample18"
>[19] "Sample19" "Sample20" "Sample21" "Sample22" "Sample23" "Sample24"
>[25] "Sample25" "Sample26" "Sample27" "Sample28" "Sample29" "Sample30"
>[31] "Sample31" "Sample32" "Sample33" "Sample34" "Sample35" "Sample36"
>
>
>I don’t know if Sample1 corresponds to inlet one of my Fluidigm chip.
>Also, I loaded three different samples onto the chip with each sample
>loaded in 6 biological replicates and 2 technical replicates. So the first
>12 inlets are one sample (BJ fibroblasts), inlets 13-24 are the second
>sample (RiPSC.BJ) and inlets 25-36 are the third sample (RiPSC.HUF1). How
>does HTqPCR know that I want the technical (and later biological)
>replicates to be combined so that I end up having three samples with each
>being represented by 6 biological replicates?
>
>
>From what and how I ask you may notice that I'm really just at the
>beginning of understanding R, HTqPCR and the coding language. Sorry for
>bugging you about this but I'd appreciate any help.
>Thanks much!
>
>Jens
>
>PS: I attached the .csv file Fluidigm gave me. There are two ways to
>export the data. One HeatMap format and one Table format. I attached both.
>
>
>On 11/13/12 5:42 AM, "Fletez-Brant, Christopher (IH/VRC) [C]"
><christopher.fletez-brant at nih.gov> wrote:
>
>>Jens,
>>
>>If your qPCRset recognizes that there are 47 features and 3 samples, then
>>you should be able to perform downstream analyses.  If you are having
>>trouble modifying sample names, I have been storing sample names in a
>>file, ordered as the actual samples are in the Fluidigm output file, then
>>assigning sample names.  That is, something like:
>>
>>Temp <- read.csv("sample.names.csv")
>>sampleNames(raw) <- Temp
>>
>>
>>Best,
>>Kipper
>>
>>On 11/11/12 3:27 AM, "Jens Durruthy-Durruthy" <jensdd at stanford.edu>
>>wrote:
>>
>>>Hi all,
>>>
>>>I'm fairly new to the HTqPCR package and to R.
>>>I wanted to analyze my Ct values that I get from the BioMark output in
>>>form
>>>of a .cvs file. I managed to read my file in with:
>>>
>>>> raw<-readCtData(files="test2.csv",format="BioMark",n.features=47,
>>>+ n.data=3,samples=samples)
>>>Warning message:
>>>In readCtData(files = "test2.csv", format = "BioMark", n.features = 47,
>>>:
>>>  Not enough sample names provided; using Sample1, Sample2, Š Instead
>>>
>>>> show(raw)
>>>An object of class "qPCRset"
>>>Size:  47 features, 3 samples
>>>Feature types:
>>>Feature names:  HSP90AB1 CD13 COL1A1 ...
>>>Feature classes:
>>>Feature categories:  OK
>>>Sample names:  Sample1 Sample2 Sample3 ...
>>>
>>>I ran 3 samples (three different cell types), each in 6 biological and 2
>>>technical replicates (12 total) but I don't know how to modify or edit
>>>my
>>>qPCRset object in order to visualize them or perform downstream analysis
>>>(PCA, clustering etc.). I read through the PDF "HTqPCR - highthroughput
>>>qPCR analysis in R and Bioconductor" by Heidi Dvinge but couldn't find
>>>an
>>>answer to my specific problem (mainly because I'm fairly new to the
>>>language
>>>R I think).
>>>
>>>If anyone could help out I would greatly appreciate it. Let me know if I
>>>need to provide additional information about the input .cvs file)
>>>
>>>Thanks!
>>>Jens
>>>
>>>:: Jens Durruthy-Durruthy ::
>>>:: Research Scholar ::
>>>:: Reijo Pera Lab ::
>>>
>>>:: Stanford University School of Medicine ::
>>>:: Institute for Stem Cell Biology & Regenerative Medicine ::
>>>:: Lorry Lokey Stem Cell Research Building ::
>>>:: 265 Campus Drive, Rm 3015 ::
>>>:: Stanford, CA 94305 € United States ::
>>>:: Mail: durruthy at stanford.edu ::
>>>:: Phone: +1-650-498-7303 ::
>>>:: Fax: 650-725-6910 ::
>>>
>>>
>>>This e-mail may contain confidential and/or privileged
>>>i...{{dropped:12}}
>>>
>>
>